text
stringlengths 296
254k
|
---|
<s>[INST] That's express VPN dot com slash leg spot. Finally the show is presented by cash app, the number one finance app in the app store. When you get it to use code lex podcast cash up unless you send money to friends Buy Bitcoin and invest in the stock market with as little as $1. I'm thinking of doing more conversations with folks who work in and around the Cryptocurrency space similar to a I there are a lot of Charlton's in the space but there are also a lot of free thinkers and technical geniuses that are worth exploring in depth and with care if I make mistakes and guest selection and details and conversation, I'll keep trying to improve, correct where I can and also keep following my curiosity wherever it takes me. So again, if you get cash out from the app store, google play and use collects podcast You get $10 in cash up. Will also donate $10 for first an organization that is helping to advance robotics and stem education for young minds around the world and now here's my conversation with Yaron brook. Let me ask the biggest possible question first. What are the principles of a life well lived. [/INST] I think it's to live with uh with thought that is to live a rational life to to think it through. I think so many people uh in a sense zombies out there, they're alive but they're not really alive because their mind is not focused, their mind is not, you know, focused on what do I need to do in order to live a great life. So too many people just go through the motions of living rather than really embrace life. So I think the secret to living a great life is to take it seriously and what it means to take it seriously is to use the one tool that makes us human. The one tool that provides us with all the values that we have our mind a reason and to use it, apply it to living people apply it to their work, they apply it to their math problems to science, to programming. But imagine if they use that same energy, that same focus that same concentration to actually living life and choosing values that they should pursue. That would that would change the world and it would change their lives. </s><s>[INST] Yeah. Actually, you know, I wear this silly suit and tie, it symbolizes to me always, it makes me feel like I'm taking the moment really seriously. [/INST] I think that's really that's right. And and each one of us has different ways to kind of uh condition of consciousness. I'm serious now and and for you it's a student. I it's a it's a conditioning of your consciousness to now, I'm focused now, I'm at work now, I'm doing my thing and I think that's that's terrific. And I wish everybody took that look. I mean, it's a cliche, but we only live once every minute of your life. You never can never live again. This is really valuable. And when people, people don't have that deep respect for their own life, their own time for their own mind. And if they did again, you know, one could only imagine look at how productive people are, look at the amazing things they produce and they do in their work, and if they applied that to everything, </s><s>[INST] wow. So, you kind of talk about reason, where does the kind of existentialist idea of experience may be fully experiencing all the moments versus fully thinking through. Is there an interesting line to separate the two? Like, why such an emphasis on reason for a life well lived versus just enjoy, like, experience? Because I think experience [/INST] in a sense is the easy part. I'm not saying it's it's it's how we experience the life that we live. And yes, I'm all with the take time to to to value what you value, but I think I don't think that's the problem of people out there. I don't think the problem is they're not taking time to appreciate where they are and what they do. I think it's that they don't use their mind in this one respect in planning their life in thinking about how to live. So the focus is on reason is because it's our only source of knowledge. There's no other source of knowledge we don't know anything with. You know, that does not come from our senses in our in our mind, the integration of the of the evidence of our senses. Now we know stuff about ourselves and I think it's important to know what itself through introspection and I count consider that part of reasoning is that is that is the introspect but I think reason is undervalued, which is funny to say because it's a means of survival. It's how human beings survive. We cannot see this is where I disagree with so many scientists and and people like Sam Harris, you mentioned the same house before the show. Um We're not programmed to know how to hunt. We're not programmed to do agriculture, we're not programmed to build computers and build networks on which we can podcast and do our shows. All of that requires effort. It requires focus, it requires energy and it requires will it requires somebody to will it, it requires somebody to choose it. And once you make that choice you have to engage that choice means that you're choosing to engage your reason in discovery and integration and then in work to change the world in which we live and you know human beings had to discover figure out solve the problem of hunting hunting. You know, everybody thinks, oh that's easy. I've seen the movie but human beings have to figure out how to do it right? You you you can't run down a bison and bite into it right? You're not gonna catch it, you're not getting, you have no thanks to bite into it. You have to build weapons, you have to build tools, you have to create traps, you have to have a strategy. All of that are quite reason. So the most important thing that allows human beings to survive and to thrive in every value from the most simple to the most sophisticated for the most material to I believe the most spiritual requires thinking so stopping and appreciating the moment is is something that I think is relatively easy once you have a plan, Once you've started through, once you know what your values are, there is a mistake people make, they attain their values and they just and they just, they don't take a moment to savor that and to appreciate that and to even pat themselves on the back that they did it right. But that's not what's screwing up the world but screwing up the world is that people have the wrong values and they don't think about them and they don't really focus on them and they don't have a plan for their own life and how to live it. </s> |
<s>[INST] Yeah, it does something um Alex Garland director who did, who wrote directed X mark in a I think about ai movies and the more care you take in making it accurate, the more compelling the story becomes smile. I'm not, I'm not sure what that is, maybe because it becomes more real to the people writing the story, maybe it just makes you a better writer. [/INST] The key to any storytelling is getting the readers to suspend their, their disbelief and there's all kinds of triggers and little tells that can break that. Um and once it's broken, it's really hard to get it back. Uh you know, a lot, a lot of times, that's the end, somebody will just close the book and not pick it up. </s><s>[INST] Uh I gotta ask you, you've answered this question, but I gotta ask you the most impossible question for an author to answer, but Which Neal Stephenson Book Should one Read 1st. [/INST] So when people ask me that, I usually ask them what they like to read, right? Because I mean there's The best known one is probably snow crash, but that's a a cyberpunk novel that's at the same time making fun of cyberpunk. Um So it's kind of got some layers to it that uh might not, it seems so funny if you don't have that, if you don't get the joke right? So um there's uh I've written, as you point out, I've written historical novels, some people like those, some people prefer those. So if that's what you like, then Krypton Omar khan or the Baroque Cycle is where you would start if you like sort of techno thrillers that are set in a modern day setting, but aren't science fiction eat per se than reme de um is one of those. And termination shock is definitely one of those. Um So it just depends on on uh what people like </s><s>[INST] what what uh when people a long time ago recommend that read Snow Crash, they said uh it's uh it's new students in light, it's the uh like if you don't want to be overwhelmed by the depth like the rigor book, like that's a good that's a good introduction to the man. So essentially you're broken down by topics, but if you wanted to read all of them, what's a good introduction to the to the man, because obviously these worlds are very different. The philosophies are very different. Yeah, What's a good introduction to the human [/INST] um </s><s>[INST] people ask the same thing with dusty esky people. It's a hard one to answer. [/INST] Maybe seven eves because it's got big themes. Um it's you know, it's about heavy heavy things happening to the human race. Um, but hopefully the story is told through a cast of characters that people can relate to. You know, it moves along. Uh so uh it does go kind of deep eventually on how rockets work and orbital mechanics and all that stuff. But um people were able to get through it anyway or some people just skip over that. It's fine, you know. </s><s>[INST] Um as an author, let me ask you what books had a big impact on your life that you have read. Is there any that jumped to mind that you learned from as a writer, as a philosopher, as a mathematician, as an engineer. [/INST] This is one of these questions where I always black out and then when I'm walking out the door I'll remember 12. </s><s>[INST] This is a random selection that doesn't represent the top [/INST] the top ones. Um Well I mentioned you know gulag archipelago that's kind of a hefty and dark but </s><s>[INST] and then it has a personal connection as well. [/INST] Yeah. Just because </s><s>[INST] like where you found the book to part time in your life where you found it, who recommended that's also part of the [/INST] story. So there's definitely that there's yeah, I circled back to Moby Dick a lot um because we read it in a really great english class I had in high school and I came in with an oppositional stance because I thought that the teacher was going to try to talk me into having all kinds of highfalutin ideas about allegory and what does this mean? What's the symbolism? And it turned out that uh it turned out to be a lot more interesting and satisfying than that. </s><s>[INST] Um what was the first powerful book? Remember reading that? Like convince you that this form could have depth. Was it Moby Dick? Was it like in high school, I'm [/INST] trying to remember, Well Moby Dick was definitely a big one. Um I used to read a lot of classics comics when I was, I don't know if you've seen these, it's a whole series of comic books that it was viral. You could uh in the, in the back of each comic book was an order form, you could check some boxes and fill out your address and mail it in more would show up. But it was like they would do the Count of Monte cristo, You know, Moby Dick, you know, robert louis Stevenson, Robinson Crusoe, you know, all the sort of classic books uh were they had put into comic book form, </s><s>[INST] it's amazing, [/INST] yeah. Reading Moby Dick if you're nine years old is a tall order, there's some very complicated sentences in there um and a lot of digressions, but if you're just looking at the comic books, like holy sh it, look at that whale, you know, um and um </s><s>[INST] and ultimately the power of the story doesn't need the complicated words, it's all about the man and the and the whale. [/INST] Yeah, yeah, so you could get kind of a grounding in a lot of classic works of literature without actually reading them, which is, you know, it's great when you're nine years old. So, so I read a lot of that stuff for sure, the annotated Sherlock Holmes, um </s> |
<s>[INST] Travis stevens actually said something that changed the way c training. He said it as a side comment, but he said that at the end of the training at the end of a good training session, your mind should be exhausted not your body. And um I've for most of my life saw good training sessions where my body was exhausted. [/INST] Yes, I believe that's the case with most people. Yeah, you should come out of the training session with your mind buzzing with ideas like possibilities for tomorrow. And by the way, on that note I would go further and say that the training session doesn't finish when your body stops moving, it finishes when your mind stops moving and your mind shouldn't stop moving after that session, there should be analysis. What did I do? Well, what did I do badly, how could I do better with the things that I did? Well </s><s>[INST] can I ask you about something that I truly enjoy and I think is really powerful but most people don't seem to believe in that. But it's drilling. I don't know, maybe people are different, but I love the idea Maybe even outside of digits of doing the same thing over and over. It's like Jiro dreams of sushi, I love doing the thing Uh that nobody wants to do in doing it 10 times 100 times 1000 times more than what nobody wants to do. Um so I'm a huge fan of drilling, obviously I'm not a professional athlete, but I feel like if I actually gave myself, If I wanted to be really good at Jiu Jitsu, like reached the level of being in the top 25 when I was much younger, like really strive uh I think I could achieve it by drilling. That's, I had this belief untested. Can you challenge this idea or, or agree? [/INST] First off, fascinating. However, we're going to have to disagree. No, no, we're just going to have to start to understand. What are we talking about when we talk about drilling? It's a very vague term, Okay, if you're um at this moment, many of your listeners are probably seeing having the same thought process, which is a drilling. Yeah, I know what that is. We go into the gym and we pick a move and we practice it for a certain number of repetitions and if I do that, I'm going to get better at the technique. Okay, um they're wrong, we've got to have a much more in depth understanding of what the hell we're talking about when we talk about drilling, ultimately any movement in the gym, it doesn't improve the skills you already have or build new skills is a waste of time, A waste of resources. Everything you do should be done with the aim and the understanding that this is going to make me better at the sport. I practice if it's not, shouldn't be there. The majority of what passes for drilling and most training halls will not make you better. Including some of the most cherished forms of drilling, which is repetition for numbers. The moment you say to someone, I want you to do this 100 times. What are they really thinking about volume? They're saying, Okay, I'm at repetition, 78 I'm at 80, 20 more to go or they're talking their primary thought process is on numbers. That's not the point of drilling. The point is skill acquisition when people drill, don't get them focused on numbers, get them focused on mechanics, that's what they have to worry about. I never have my students drill for the numbers ever. This one </s><s>[INST] crew, get the funk outta [/INST] Here, are you kidding me? You're gonna get better with that. Okay, get them working on the sense of gaining knowledge. That's my job. I have to give them knowledge. I have to explain to them what they're trying to do. That starts them on the right track. But knowledge is one thing, skill is another. If judas was just about knowledge then All the 60 and seven year old red balance would be the world champions. They're not Judicial isn't won by knowledge. Its one by Skill. Knowledge is the first step in building skill. So my job as a coach is to transmit knowledge, then I have to create training programs with a path from knowledge to polished skill is carried out. That's the interface between me and my students. And so I give them drills where the whole emphasis is upon getting a sense where they understand what are the problems they're trying to solve and working towards practical solutions. They never work with numbers. They work with mechanics and feel. Then you have to bring in the idea of progression. When you drill, there's zero resistance. When you fight in competition, there's 100% resistance. You can't go from 0 to 100. There has to be progress over time Where I have them work and drills with slightly increasing increments of resistance. And just as we talked about earlier with the weightlifter who doesn't start with £500 but who begins with the bar And then over time builds the skills that one day out there in the future he will live £500. So to that Judy Tommy that you're working on today is feeble and pathetic. But five years from now you'll win a world championship with it. Mm you can't have this naive idea of drilling is something you just come out. Mm hmm randomly pick a move and you work for numbers until you're satisfied a certain set of numbers that your coach threw at you and then you think you're going to get better. There's even dangerous with drilling, There is no performance increase that comes once you get to a certain level and you just keep doing the same damn thing, let's say for example, You come out and you hit 100 repetitions of the armed guard duty guitar me from guard position. And you're all proud of yourself because you hit 100 repetitions and your body's tired and you're telling yourself, man, I got a good workout and you come in tomorrow, you do exactly the same thing. You come in the day after that and a week goes by and you've done the same thing. Then a year later you do the same thing. Ask yourself, has your Duty Guitar me really gotten better? No, you've performed literally thousands and thousands of repetitions. You've spent an enormous amount of training time and energy that could have gone in different directions on something which didn't make you any better drills have diminishing returns. Once you get to a certain skill level, If you just keep hammering on the same thing in the same fashion for the same amount of time, you stop getting better. </s> |
<s>[INST] Well, I was referring more on the like, not an actual voice. I meant like there's some kind of like, subjective experience, feels like it's it's fundamentally about storytelling to ourselves. It feels like like the feeling is a story that is much uh much simpler abstraction than the raw sensor information, so that it feels like it's a very high level abstraction that uh is useful for me to feel like entity in this world. Most useful aspect of it is that because I'm conscious, I think there's an intricate connection to me, not one wanting to die. So like it's a useful hack to really prioritize not dying, like those seem to be somewhat connected. So I'm telling the story of like it's richly feels like something to be me and the fact that me exists in this world, I want to preserve me and so that makes it a useful agent hack. [/INST] So I will just refer maybe to the first part as he said, about the kind of story of describing who you are. Um I was uh thinking about it even so, you know, obviously, I I like thinking about consciousness. I like thinking about Ai as well and I'm trying to see analogies of these things in a I what would it correspond to? So uh I am, you know, open the train uh a model called GPT which can generate uh pretty amusing text on arbitrary topic. And um and one way to control GPT is by putting into prefix at the beginning of the text some information. What would be the story about uh you can have even chat with uh you know with GPT by saying that the chat is with legs or Elon musk or so. And GPT would just pretend to be you or Elon mask or so and and it almost feels that this uh story that we give ourselves to describe our life. It's almost like a things that you put into context of GPT </s><s>[INST] the Primaries, the but the context we provide. The GPT is uh is multimodal. [/INST] It's so GPT itself is multi model. GPT itself hasn't learned actually from experience of single human, but from the experience of humanity, it's a chameleon, you can turn it into anything and in some sense by providing context uh you know, behaves us the thing that you wanted it to be. And it's interesting that the, you know, people have the stories of who they are and as you said, these stories, they help them to operate in the world. But it's also, you know, interesting guess various people find it out through meditation or so that uh there might be some patterns that you have learned when you were a kid that actually are not serving you anymore. And he also might be thinking that that's who you are and that's actually just a story. </s><s>[INST] Yeah, so it's a useful hack, but sometimes it gets us into trouble. It's a local optima, you wrote that Stephen Hawking, He tweeted Stephen Hawking asked, What breathes fire into equations, which meant, what makes given mathematical equations realize the physics of the universe similarly, I wonder what breathes fire into computation? What makes given computation conscious? Okay, so how do we engineer consciousness? How do you breathe fire and magic into the machine? [/INST] So, um it seems clear to me that not every computation is conscious. I mean, you can let's say just keep on multiplying one matrix over and over again and might be gigantic. matrix. You can put a lot of computation. I don't think it would be conscious. So in some sense, the question is, what are the computations? Which could be conscious? I mean. So one assumption is that it has to do purely with computation that you can abstract away matter. Other possibility is that it's very important was the realization of computation that it has to do with some uh four skills or so. And they bring consciousness at the moment. My intuition is that it can be fully abstracted away. So in case of computation you can ask yourself what are the mathematical objects or so that could bring such a properties. So for instance, if we think about the models uh ai models that what they truly try to do or like a models like GPT is uh uh huh. You know, they try to predict the next word or so. And this turns out to be equivalent to uh compressing text. Um And because in some sense, compression means that you learn the model of reality and you have just to uh remember where are your mistakes? The battery are in predicting the and and and in some sense when we look at our experience. Also when you look for instance at the car driving, you know in which direction it will go. You're good like in prediction. And you know, it might be the case that the consciousness is intertwined with compression. It might be also the case that self consciousness uh has to do with compressor trying to compress itself. So uh okay. I was just wondering what are the objects in you know mathematics or computer science which are mysterious that could uh that that that could have to do with consciousness. And then I thought uh you know you you see in mathematics there is something called God lt R. M. Which means okay you have if you have sufficiently complicated mathematical system, it is possible to point the mathematical system back on itself. In computer science there is uh something called helping problem. It's somewhat similar construction. So I thought that you know if we believe that that and the assumption that consciousness has to do with uh with compression uh then you could imagine that the that the as they are keep on compressing things then at some point it actually makes sense for the compressor to compress itself. </s> |
<s>[INST] know I'm a computer science guy so I have some intuition but it's the cyber world it's really hard to build up an intuition what is safe and not you know I've seen a lot of people you know just logging out of your devices all the time like regularly just like that physical access stuff is a lot of people don't take. I can just like walk in into the offices of a lot of C. E. O. S. It's like everything is wide open uh for for physical access of those systems which is kind of incredible for uh somebody that sounds really shady but it's not I've written key loggers like things that record everything you type in the policy mood and like I I did that for um during my PhD I was recording everything you do on your device and everything you do on your computer to understand. Like people sign up to the study, they willingly do this to understand behavior. I was trying to use machine learning to identify who you are based on different biometric and behavioral things which allows me to study human behavior and to see which is uniquely identifiable. And the goal there was to remove the need for a password but how easy it is to write a thing that logs everything you type. I was like wait a minute. Like I can probably get a lot of people in the world to run this for me. I can then get all of their passwords. I mean you could do so much like I can I can run the entirety of the C. I. A. From just myself if I was and I imagine there's a lot of really good hackers like that out there, much better than me. Um So I tried to prevent myself from being all the different low hanging fruit attack vectors in my life. I tried to make it difficult uh To be that but then I'm also aware that there's probably people that are like five steps ahead [/INST] you're you're doing the right thing. What what I always advocate is the low hanging fruit is what keeps you from being a target of opportunity because you're half assed hackers, you're lazy hackers, you're unskilled hackers, They're looking for low hanging fruit. They're looking for the person who gets the Nigeria email about how you could be getting $5 million. If you just give me your bank account, that's what they're looking for. And the thing that's scary is that if you're not a target of opportunity, if you become a intentional target then there's almost nothing you can do because once you become an intentional target then your security apparatus, they will and they will create a dedicated customized way vector of attacking your specific security apparatus. And because security is always after, right? There's always there's the there's the leading advantage and the trailing advantage when it comes to attacks, the leader always has the advantage because they have to create the attack before anybody else can create a way to protect against the attack. So the attack always comes first and that means they always have the advantage. You are always stuck just leaning on, this is the best security that I that I know of. Meanwhile there's always somebody who can create a way of attacking the best security out there and once they win, they have a monopoly. They have all that time until a new defensive countermeasure is deployed. </s><s>[INST] Yeah, I tend to think exactly as you said that the long hanging fruit protects against like yeah, crimes of opportunity. And then I assume that people can just hack in if they really want think [/INST] about how much anxiety we would be able to solve if everybody just accepted that. </s><s>[INST] Well there's several things you do first of all I to be honest, it just makes me, it keeps me honest, not to be a douche bag or like uh knot yeah, to to uh, to assume everything could be public and you know, and so don't trade information that could hurt people. It was made public. So I tried to do that and the thing I try to make sure is I like home alone style try, [/INST] I </s><s>[INST] really would like to know if I was hacked [/INST] right. And </s><s>[INST] so I try to assume that I will be hacked and detected have [/INST] a trip wire or something. </s><s>[INST] Yeah trip wires through through everything. Uh, and not paranoid trip wires just like open door. But that's, that's, I think that's probably the future of life on this earth is you're going like everybody of interest is going to be hacked um, that hopefully inspires. No, this is outside of company. There's um, these are individuals. I mean there's of course if you're actually operating like I'm just, who am I? I'm just a scientist person, um podcasting person. So if I was actually running a company or was an integral part of some kind of military operation, then you have to probably have to have an entire team that's now doing that battle of like being trying to be ahead of like the best hackers in the world that are attacking, but that that requires a team that like full time is their focus and then you still get in trouble, [/INST] correct? Yeah. So what I've seen as the norm, well what I've seen is the cutting edge standard for corporations and the ultra with the ultra wealthy and even uh intelligence organizations is that we have trip wires. It's better if you can't prevent from being hacked. The next best thing is to know as soon as you get hacked because then you can essentially terminate all the information. If you know it fast enough, you can just destroy the information. This is what the ultrawealthy do they have multiple phones. So as soon as one phone gets hacked, the trip wire goes off, the operating system is totally deleted along with all data on the phone. And a second phone is turned on with a whole new separate set of metadata and now they for them there's no break in service. It's just up this phone went black. It's got a warning on it that says it was hacked so trash it because they don't care about the price of the phone, pick up the next phone and we move on. That's that's the best thing that you can do essentially outside of trying to out hack the hackers. Uh and then even in your intelligence and military world where the cyber war cyber warfare is active. The people who are aggressing are not trying to create aggression that beats security. They're trying to find aggressive techniques, offensive techniques that have no security built around them yet because it's too cost and time intensive to protect against what you know is coming. It's so much more efficient and cost effective to go after new vectors. So it just becomes like, uh, it becomes almost a silly game of of your neighbor gets a guard dog. So you get a bigger guard dog and then your neighbor gets a fence. So you just constantly outdoing each other. It's called the security paradigm. People just, they just one up each other because it's never worth it to just get to the same level. You're always trying to outdo each other. </s> |
<s>[INST] Tell you how to be curious about the world and how to be open minded. [/INST] Yeah. </s><s>[INST] Last question, what role does love play in the human condition? Since we've been talking about love and robots? Uh and you're fascinated by social robotics. It feels like all of that operates in the landscape of something that we can call love. [/INST] Love. Yeah, I think there are a lot of different kinds of love. I feel like it's we need like the Eskimos have all these different words for snow, We need, we need more words to describe different types and kinds of love that we experience. But I think love is so important. And I also think it's not zero sum. That's the that's the really interesting thing about love is that, you know, I had one kid and I loved loved my first kid more than anything else in the world. And I was like, how can I have a second kid and then love that kid. Also, I'm never going to love it as much as the first, but I love them both equally. It's just like my heart expanded and so I think that people who are threatened by love towards artificial agents, they don't need to be threatened for that reason, </s><s>[INST] artificial agents will just if done right will just expand your capacity for love. I [/INST] think so, </s><s>[INST] I agree beautifully. Put, Okay, this is awesome. I still didn't talk about half the things I want to talk about, but we're already like way over three hours. So thank you so much. I really appreciate you talking today. You're awesome. You're an amazing human being, a great roboticist, great writer. Now it's an honor that you would talk with me. Thanks for doing it [/INST] right back at you. Thank you. </s><s>[INST] Thanks for listening to this conversation with kate darling. To support this podcast. Please check out our sponsors in the description and now let me leave you with some words from Maya Angelou. Courage is the most important of all the virtues because without courage, you can't practice any other virtue consistently. Thank you for listening and hope to see you next time [/INST] </s> |
<s>[INST] Yeah, exactly transforms it fundamentally. Because say that machine does tell you that it can predict perfectly. And yet there is this deep experience of free will and then that that changes the question completely. And it starts actually getting to the question of uh the the A. G. I. The touring questions of the demonstration of free will. The demonstration of intelligence, the demonstration of consciousness, does that equal consciousness, intelligence and free will. [/INST] But see uh lex if if every time I was contemplating a decision, you know, this machine had printed out an envelope, you know, where I could open it and see that it knew my decision. I think that actually would change my subjective experience of making decisions. You know, I mean, </s><s>[INST] there's knowledge change our subjective [/INST] experience. Well, you know, I mean, I mean the knowledge that this machine had predicted everything. I would do, I mean, it might drive me completely insane, right? But at any rate, it would change my experience to, you know, to not just discuss such a machine as a thought experiment, but to actually say it. </s><s>[INST] Yeah, [/INST] I mean, I mean, you know, you could say at that point, you know, you could say, you know, why not simply call this machine a second instance creation of me and be done with it. Right? What? We know what what, why even privilege the original me over this perfect duplicate that exists in the machine. </s><s>[INST] Or there could be a religious experience with it too. It's kind of what God throughout the generations is supposed to have, that God kind of represents that perfect machine is able to um I guess actually, Well, I I don't even know what what what are the religious interpretations of free will uh does. So if God knows perfectly everything in religion, in the various religions, where does free will fit [/INST] into that, Do you know? That has been one of the big things that theologians have argued about for thousands of years. So, you know, I'm I'm not a theologian, so maybe I shouldn't go there. </s><s>[INST] There's not a clear answer in a book, like, [/INST] uh I mean, I mean, this is, you know, the calvinists debated this. The, you know, this has been, you know, I mean, different religious movements have taken different positions on that question, but that is how they think about it, you know, meanwhile, you know, a large part of sort of what, what animates, you know, theoretical computer science, because you know, we're asking sort of what are the ultimate limits of, you know, what you can know or you know, calculator figure out by, you know, entities that you can actually build in the physical world. Right? And uh if I were trying to explain it to a theologian, maybe I would say, you know, we are studying, you know, to what extent, you know, gods can be made manifest in the physical world. I'm not sure my colleagues </s><s>[INST] would like that. So let's talk about quantum computers. [/INST] Sure, sure. </s><s>[INST] As you said, quantum computing at least in the 19 nineties was a profound story at the intersection of computer science, physics engineering, math and philosophy, so that there's this broad and deep aspect to quantum computing that represents more than just the quantum computer. But can we start at the very basics, what is quantum computing? [/INST] Yeah, so, it's a proposal for a new type of computation or let's say, a new way to harness nature to do computation that is based on the principles of quantum mechanics. Okay, now, the principles of quantum mechanics have been in place since 1926, you know, they haven't changed. Uh you know, what's new is, you know, how we want to use them. Okay, so what does quantum mechanics say about the world? You know, the the physicists, I think over the generations, you know, convinced people that that is an unbelievably complicated question and you know, just give up on trying to understand it. I can let you in, not not being a physicist, I can let you in on a secret, which is that it becomes a lot simpler if you do what we do and quantum information theory and sort of take the physics out of it. So, uh the way that we think about quantum mechanics is sort of as a generalization of the rules of probability themselves. So, um you know, you might say there's a, you know, there was a 30% chance that it was going to snow today or something. You would never say that there was a negative 30% chance, right, that would be nonsense much less. Would you say that there was, you know, an eye percent chance, you know, square root of minus 1% chance? Uh Now, the central discovery that sort of quantum mechanics uh made is that uh fundamentally the world is described by uh you know, these sort of, let's say the possibilities for for, you know, what a system could be doing are described using numbers called amplitudes, Okay. Which are like probabilities in some ways, but they are not probabilities, they can be positive for one thing that can be positive or negative. In fact, they can even be complex numbers. Okay. And if you've heard of a quantum superposition, this just means that uh some state of affairs where you assign an amplitude one of these complex numbers to every possible uh configuration that you can see a system in on measuring it. So for example, you might say that uh an electron has some amplitude for being here and some other amplitude for being there right now, if you look to see where it is, you will localize it right? You will sort of force the amplitudes to be converted into probabilities that happens by taking their squared absolute value. Okay. And then, and uh and then, you know, you can say either the electron will be here or it will be there, you know, knowing the amplitudes, you can predict the at least the probabilities that it will that you'll see each possible outcome. Okay. But while a system is isolated from the whole rest of the universe, the rest of its environment, the amplitudes can change in time by rules that are different from the the normal rules of probability and that are, you know, alien to our everyday experience. So any time anyone ever tells you anything about the weirdness of the quantum world, you know, assuming that they're not lying to you, right? They are telling you, you know, and yet another consequence of nature being described by these amplitudes. So most famously what amplitudes can do is that they can interfere with each other. Okay, so uh in the famous double slit experiment, what happens is that you shoot a particle like an electron, let's say, at a screen with two slits in it. And you find that there are, you know, on a second screen. Now there are certain places where that electron will never end up, you know after uh it passes through the first screen and yet if I close off one of the slits then the electron can appear in that place. Okay, So so by decreasing the number of paths that the electron could take to get somewhere, you can increase the chance that it gets there. Okay, now, how is that possible? Well, it's because, you know, as we would say now the electron has a superposition state. Okay, it has some amplitude for reaching this point By going through the first split it has some other amplitude for reaching it by going through the second slit but now if one amplitude is positive and the other one is negative then though, you know, I have to add them all up. Right, I have to add the amplitudes for every path that the electron could have taken to reach this point and those amplitudes, if they're pointing in different directions, they can cancel each other out. That would mean the total amplitude is zero. And the thing never happens at all. I close off one of the possibilities, then the amplitude is positive or it's negative. And now the thing can happen. Okay, so that is sort of the one trick of quantum mechanics and now I can tell you what a quantum computer is. Okay, a quantum computer is a a computer that tries to exploit you know these exactly these phenomena superposition amplitudes and interference in order to solve certain problems much faster than we know how to solve them otherwise. So is the basic building block of a quantum computer is what we call a quantum bit or a qubit. That just means a bit that has some amplitude for being zero and some other amplitude for being one. So it's a superposition of zero in one states. Right? But now the key point is that if I've got let's say 1000 cubits, the rules of quantum mechanics are completely unequivocal that I do not just need one and you know, I don't just need amplitudes for each cube it separately. Okay. In general I need an amplitude for every possible setting of all 1000 of those bits. Okay, so that what that means is two to the 1000 power amplitude. It's okay if I if I had to write those down, Let's say in the memory of a conventional computer, if I had to write down two to the 1000 complex numbers That would not fit within the entire observable universe. Okay. And yet, you know quantum mechanics is unequivocal that if these cubits can all interact with each other and in some sense I need to to the 1000 parameters amplitudes to describe what is going on. Now, you know, now I can tell you know where all the popular articles, you know about quantum computer and go off the rails is that they say, you know, they they sort of sort of say what I just said and then they say oh so the way a quantum computer works is just by trying every possible answer in parallel, You know, you know that that sounds too good to be true and unfortunately it kind of is too good to be true. The problem is I could make a superposition over every possible answer to my problem, you know, even if there are two to the 1000 of them right, I can I can easily do that. The trouble is for a computer to be useful. You've got at some point you've got to look at it and see and see an output. Right? And if I just measure a superposition over every possible answer, then the rules of quantum mechanics tell me that all I'll see will be a random answer. You know, if I just wanted a random answer. Well, I could have picked one myself with a lot less trouble. Right? </s> |
<s>[INST] can't. I mean, I'm gonna get teary eyed, No, but I have allergies, but also that's pretty powerful what you're saying, I certainly share your ideas. It's something I struggle with. I've, just by instinct, you should read the idiot by dostoevsky by instinct. I'm I love being naive and uh seeing the world from a hopeful perspective, from an optimistic perspective, it is it's sad that that is something you pay a price for in this world, like in the academic world, especially as you're coming up uh through through schooling, but just actually, it's a hit on your reputation throughout your life and it's a sad truth, but you have to, like for many things, if it's a principle behold you have to be willing to pay the costs. And ultimately, I believe that in part, a hopeful view will help you realize the best version of yourself, because optimism is the kind of optimism is productive, like believing that the world is and can be amazing, is allows you to create a more amazing world somehow. I mean, I'm not sure uh I'm not sure if it's a human nature of a fundamental law of physics, I don't know, but believing the impossible in the sense of being optimistic about the thing is it's it's some of like going back to what you've said is like believing that a radical, that a powerful single idea that a single individual can revolutionize some framework that we're operating in that will change the world for the better, believing that allows you to have the chance to create that. And so I'm with you on the optimism, but you may have to pay a cost of optimism and uh naive hopefulness. [/INST] I mean in some sense, optimism limits freedom. Uh, I think if we don't really have much choice in choosing what is perfect if it exists is is an ideal. Um then there isn't much room for creativity and that's a danger of optimism as someone who uh I would like to be creative, I think, I think it was Warren's, Yvonne said, uh excepting dreams, you're never really free and that's something I think about a lot. Um he's an interesting guy also, I really like him. </s><s>[INST] Uh huh on that topic, you do have uh, a bit of an appreciation in connection with music. I saw you play some guitar a few months ago, uh what what uh can you put in like a philosophical sense, your connection to music? Uh what insights about life, but just the way you see the world do you get from music? I think the [/INST] role music has played in my life was originally motivated by sort of wanting to prove things to myself, I really have no ear for music, I have a terrible sense of pitch and I think a lot of music relies on very standard teaching. If you think about uh lessons, for example, music lessons, there's there's sort of a routine to them which is so archaic and traditional that there's no room for for for deviation. I think all of that suggested to me that I would never have a relationship with music, I loved listening to music, it was just, it was difficult to me, it was sort of saddened me. Um I wanted to know if there was any way I could build a connection to music, given who I am, my own idiosyncrasies. Um what challenges I have, I decided to try to learn music theory before I touched an instrument. Uh Mhm. I think that gave me a very unique opportunity instead of spending my time fruitlessly at the beginning on the syntax of a particular instrument. This is how you this is your posture on the piano, this is how you hold your fingers. Uh I tried instead to learn what made music work. And the wonderful thing about that was I'm pretty sure that any instrument with discrete notes is mine for the taking within a day or so of having the ability to play with it. So I think approaching music abstract lee gave me the ability to install and she ate it uh everywhere. And I think it also taught me something about self teaching. Like recently I've tried getting into into classical music because at least traditionally this is the thing which is thought to require the most rigor. Um and traditional uh teaching, I think it's essentially taught me, even if I'll never be a great classical performer, that there is nothing one can't really teach themselves in this </s><s>[INST] in [/INST] this era. So I've been I've been enjoying uh whatever connection I have with with music. The other thing I'll say about it is that it's a it's a very rewarding learning process. We know, for example, that music uh accesses are neurochemicals very directly. And if you teach yourself a little bit of theory and are able to in stan she ate it on an instrument without wasting your time or spending your time tediously on learning the particulars of that instrument, you can instantly sit down and access your own dopamine loops. And so you don't really need to motive motivate yourself with music because, you know, like you're giving your brain drugs, you know, who needs motivation to to give themselves drugs and learn something. So, uh I think I think more people should be, should be playing music. I think a lot of people don't realize how easy it can be to approach if you take a sort of um on standard approach </s> |
<s>[INST] Yeah, it does seem to be, I don't know what it is, but it does seem to be that fear is a good indicator of something you should probably face. Like fear kind of shows the way a little bit. Uh not always, you might not want to go into the cage with a lion, but but maybe you should maybe uh let me ask sort of a darker question because we're talking about dostoyevsky, we might as well. Do you ah do you end connected to the hope, freeing innocent people? Do you think about mortality? Do you think about your own death? Are you afraid of death? [/INST] I'm not afraid of of death. I do think about it more now because I'm now in my mid fifties. So I used to not think about it much at all. But uh the harsh reality is that I've got more time behind me now that I do in front of me. And it kind of happens all of a sudden to you realize, wait a minute. I'm I'm actually on the back nine now. Uh so uh yeah, my mind moves to it from time to time. I don't dwell on it. I'm not afraid of it. Uh my own personal religious commitments, I'm I'm christian and my religious uh, commitments buoy me uh that uh you know that that death and I I believe this death is not not not the end. So I'm not afraid of it now. This is not to say that. I wanna I wanna I wanna rush to the afterlife. I'm good right here For a long time and I hope I've got, you know, 30, 40 more years to go. But but but no, I don't I don't really I don't fear death. Uh we're finite creatures, we're all gonna we're all gonna </s><s>[INST] die. Well, the mystery of it. Uh you know, for somebody, at least for me, we human beings want to figure everything out. Uh whatever the afterlife is there still a mystery to it? That uncertainty, it can be terrifying if you ponder it, but maybe uh what you're saying is uh you haven't pondered it too deeply so far and it's worked out pretty good, [/INST] it's worked out. No, no no complaints. </s><s>[INST] So you said again the ceo of ski kind of, well it was exceptionally good at getting to the core of what it means to be human. Do you think about like the why of why we're here? The meaning of this whole existence? [/INST] Yeah, no, I I do I think uh and actually think that's the purpose of an education, what does it mean to be a human being and in one way or another, uh we set out to answer those questions and we do it in a different way. Uh I mean some may look too uh philosophy to answer these questions. Why is it in one's personal interest to uh, to do good to do just to do justice. Uh, some may uh look at it through the economists lens. Uh, some may uh look at it through the microscope in the laboratory that the phenomenal world is uh, is the meaning uh, of life. Others may say that that's one vocabulary. That's one description, but the poet describes a reality to the same degree as a physicist. But that's the purpose of of of an education. It's to sort of work through these issues. What does it mean to be a uh, what does it mean to be a human being? And I think it's a fascinating journey and I think it's a lifelong uh endeavor to figure out what is the thing that nugget that makes us uh human. Do </s><s>[INST] you still see yourself as a student? Of course? [/INST] Yes. I mean that's uh that's the best part about going into into university teaching your, your your lifelong student. I'm always learning I learned from my students and with my students and uh my colleagues and your, you continue to read and and and learn and and modify opinions and I think it's just a wonderful thing. </s><s>[INST] Well Ron um, I'm so glad uh, that uh, somebody like you is carrying the fire of what is the best of Harvard. So it's a huge honor that you spend so much time, waste so much of your valuable time with me. I really appreciate that. I think a lot of people love it. Thank you so much for talking today. Thank you. Thanks for listening to this conversation with Ronald Sullivan and thank you to Brooklyn and sheets, wine, access online wine store, monk pack, low carb snacks and blink ist app that summarizes books, click their links to support this podcast and now let me leave you with some words from NElson Mandela. When a man has denied the right to live the life he believes in, he has no choice but to become an outlaw. Thank you for listening and hope to see you next time. [/INST] </s> |
<s>[INST] Now you're just sweet talking to me because I'm a fan, but I appreciate this. [/INST] Okay, well, we could get into what I see as being a structural flaws in the brothers caramel anyway, now </s><s>[INST] I'm offended and triggered. [/INST] Yeah, I mean, this is coming out of Macbeth and saying, I didn't think it was much good. </s><s>[INST] Yeah, there's structural flaws. [/INST] Yeah. Ending stank middle wasn't very good when I read that, that that novel, I was I was flawed by a couple of things. One is one is of course at the moment where we realize the devil appears the moment that even says to his brother, you know, he visits me and you realize that he's talking about the devil whole novel goes into this totally different space anyway, it's even more than you've already realized the novel's about And then when the conversation occurs between heaven and the devil, remember, describes him as dressed as a french dressed in the french style Of the early part of the 19th century. Strange, the devil will be dressed like that, but sort of um and and and if you remember that, he and he sort of crossed legs and rather urbane figure. But but the devil mentions in passing to even that he says, I don't know why gratitude is not a Mhm. An interesting that's been given to me, and </s><s>[INST] you're not allowed. This is not given the role of being the devil. This is not one of the things. [/INST] And you think, and of course, only a genius of scat stature could mean a lesser genius would have made a whole novel out of that insight only you can just throw it away because there's such an abundance of riches that he still has to get through structural problems aside. But the the </s><s>[INST] uh the passive aggressive nature, the microaggression in this [/INST] conversation is </s><s>[INST] palpable [/INST] little knife fight? No, but the reason I mention this because of course, when I was this is such a brilliant insight, because why why would why would gratitude not be a sentiment that the devil was capable of? The answer is of course, that if the devil was capable of gratitude, he wouldn't be the devil could be somebody else, he has to be incapable of gratitude. </s><s>[INST] Do you think for dostoevsky, that was as strong as an insight as it is for you? Because I think that's a really powerful idea that's with gratitude. You don't get the resentment that rots you from the core. [/INST] Yes. I think it was one of the just endless things that he saw in us and and the way I put it is that, I mean, I also think thinking in terms of the era of deconstruction, which is one of the things I'd like us to call the era that's now ending the era of deconstruction was the era that started, let's say from the 60s onwards and was originally an academic game that then spilled out into the wider culture, which was, let's take everything apart, Let's pull it all apart. Um, there are lots of problems with it. One is it's quite boring. You don't get an awful lot from it. Uh, you also have the problem of what Children find when they try to do this with bicycles, which is, they can take it apart quite easily. They can't put it back together. And the era of taking things apart as a game is one we've lived through and it's been highly destructive, but you can do it for quite a long time. I'm going to look at this society and I'm going to take it apart by showing systemic problems? I'm going to at the end of that. What have you got, what have you done? What have you achieved? We need to interrogate this. Okay, interrogate. By all means ask questions. Interrogators, a deliberate hostility to this. I'm going to interrogate this thing and take it apart and again at the end of it, what have you got? Whether you're interrogating a text or a piece of music or an idea or a society? Fine question endlessly question. Yes, interrogate assumes it's all um criminal in a cell and it's guilty and therefore it must be taken apart. And that's what we've been doing for decades in the west. And that's resentment. That's one byproduct of resentment. You can't build the thing, but you know how to take it apart. </s><s>[INST] Is a little bit of resentment. Good. So you have you know that I love tom waits and here's a song where a little drop of, I like my town with a little drop of poison. Is it good to do that? Is it good to have a little bit of poison in your drink? [/INST] Depends what the poison is And it depends if you know not to have another drink, it might be the case you find out of some alcoholics to that one was too many intent is not enough. </s><s>[INST] Mm. So there's a natural in this case this kind of deconstruction is a slippery slope. It becomes an addiction becomes a drug and you just can't stop, [/INST] you'd have to wean yourself off it and try to start creating again? You'd have to start trying to put things together again. Um Something I think might be in the throes of starting as it happens. </s><s>[INST] Well speaking of taking things apart and not putting them together again. The idea of critical race theory. Um can you, to me explain so I'm an engineer and have not been actually paying attention much unfortunately to these [/INST] people in your field were until it comes along and smacks you in the face. </s> |
<s>[INST] according to you. Yeah. [/INST] According to me, </s><s>[INST] I don't know if others would say the same, right? [/INST] Um </s><s>[INST] I feel like ultra competitive people, like, I've heard joe Rogan say this to me is like, he's a lot less competitive than you used to be. I don't know that. Oh, [/INST] I believe it. No, I totally believe it. Like, because as you get, you can still be like, I care about winning. Like when, you know, I play a game with my buddies online or, you know, whatever it is, polito p is my current obsession. Like, thank </s><s>[INST] you for passing on your obsession to me [/INST] now. Yeah, </s><s>[INST] I'm playing now, [/INST] but </s><s>[INST] I'm terrible and I enjoy playing terribly. I don't want to have a game because that's going to pull me into your monster of of like, competitive [/INST] play. I'm </s><s>[INST] enjoy, enjoy playing on the I can't you just [/INST] do that. You just do the points thing, you know, against the bots. </s><s>[INST] Yeah, against the bots and I can't even do the uh there's like a hard one and there's a very [/INST] crazy. Yeah, </s><s>[INST] that's crazy. I can't I don't even enjoy the hard won, the crazy. I really don't enjoy because it's intense you have to constantly try to win as opposed to enjoy building a little world and you [/INST] know there's no time for exploration, polito P. A. You gotta get well when once you graduate from the crazies then you can come play the </s><s>[INST] graduate from the crazy [/INST] in order to be able to play a decent game against like you know our group. Um You'll need to be, You need to be consistently winning like 90% of games against 15 crazy bots and you'll be able to like maybe I could I could teach you it within a day honestly. Um </s><s>[INST] how to be the crazies, [/INST] crazies and then and then you'll be ready for the big leagues </s><s>[INST] generalizes to more than just polito P. A. But. Okay. Why were we talking about [/INST] Paula? Topia </s><s>[INST] losing hurts, losing [/INST] hers. Yes, competitiveness over time. </s><s>[INST] I [/INST] think it's more that at least for me I still care about play about winning. When I choose to play something. It's just that I don't see the world as a zero sum as I used to be. You know. Um I think as you one gets older and wiser you start to see the world more as a positive something. Or at least you're more aware of externalities of of scenarios of competitive interactions. Um And so yeah I just like I'm more and more aware of my own stuff. You know like if I have a really strong emotional response to losing and that makes me feel shitty for the rest of the day and then I beat myself up mentally for it. Like, I am now more aware that that that's unnecessary negative externality. So I'm like, okay, I need to find a way to turn this down, you know, dial this down a bit. </s><s>[INST] Was poker the thing that has, if you think back in your life and think about some of the lower points of your life, like the darker places you've got in your mind, didn't have to do something with poker. Like what did losing spark the, um, the descent into darkness? Or was it something [/INST] else? I think my darkest points in poker well, when I was wanting to quit and move on to other things, but I felt like I hadn't ticked all the boxes I wanted to tick like I wanted to be the most winningest female player, which is by itself a bad goal. Um, you know, that was one of my initial goals and was like, well, I haven't, you know, and I wanted to win a WPT event, I've won one of these and one of these, but I want one of those as well. And that sort of, again, like it's a drive of like over optimization to random metrics that I decided were important um, without much wisdom at the time, but then like carried on. Um, that made me continue chasing it longer than I still actually had the passion to chase it for and I don't, I don't have any regrets that, you know, I played for as long as I did because who knows? You know, I wouldn't be sitting here, I wouldn't be living this incredible life that I'm living now. Um </s><s>[INST] this is, this is the height of your life right now, [/INST] This is it experience absolute pinnacle here in your robot land. </s><s>[INST] Yeah, [/INST] creepy light. It is, I mean I, I wouldn't change a thing about my life right now and I feel very blessed to say that. Um so, but the dark times were in the sort of like 2016 to 18 even sooner really, where I was like, I stopped loving the game and I was going through the motions and I would say that and, and then I was like, you know, I would take, the loss is harder than I needed to because I'm like, it's another one. And it was, I was aware that like, I felt like my life was ticking away and I was like, is this gonna be what's on my tombstone? Oh yeah, she played the game of, you know, this zero sum game of poker slightly more optimally than her next opponent, Like cool, great legacy, you know, so I just wanted, you know, there was something in me that knew I needed to be doing something more directly impactful, um and just meaningful, It was like a search for meaning. And I think it's a thing, a lot of poker players, even a lot of, I imagine any games players who sort of love intellectual pursuits, um, you know, I think you should ask magnus Carlsen this question. Yeah. Like it must be so hard for him. You know, he's been on the top for so long and it's like, well now what he's got this incredible brain, like what to put it to. Um, and yeah, it's, it's </s> |
<s>[INST] Yeah. [/INST] What they do is because they know that when someone gets a cancer diagnosis, that their ability to comprehend information changes. When you get a cancer diagnosis, you all the tropes they happen, you're hearing it's weird, you're basically having like an anxiety attack. If I had to guess it's like a hardcore anxiety attack. And then, you know, a nurse is there with me as he's explaining it. And then her job is even though he's telling me how to get to the machine that's gonna scan my body to see if it's gotten into my brain, uh he knows I'm not gonna remember that. And so this nurse, when you're in this like fog takes you at least took me to the machine that does the scan, You're not going to get that data back for a few days. And so that's where you really live in the real world. That's the real </s><s>[INST] world. Such a fascinating moment and the days that follow. And even at that moment because that doctor you talk about the matrix where like the pills and so on. You get the blue pill and the red pill. [/INST] Yeah, this </s><s>[INST] is like the, [/INST] like </s><s>[INST] the real world introduction, the human introduction is the truth. You've now just taken the red pill. You get to see the truth of reality. And here's a busy doctor just telling you like all those dreams, you've had all those illusions, you've built up to somehow your work as a comedian and actor will make you live forever somehow. It's just the basic illusion we have that we're this is this whole project. Uh it's gonna be an infinite sequence of fun things that we're gonna get to do. It's like, holy sh it's not [/INST] that's right, it's over, that's right. And there's very sophisticated ways of doing that. And there's very dumb ways of doing that. And I had really been doing a dumb way of doing that. Like I've been playing around with this idiot notion of subjective consciousness, so like, like I've been sort of kicking around this, like, I think they call it solipsism, it's like you're like, okay, I know I'm self aware, but no one else can prove that they're self aware. Like, I don't I have no way of proving that everything around me isn't just a video game isn't just some projection isn't, you know, who knows what? So maybe everybody else dies there, nBcs, but I don't because I'm the only thing I know that has subjective consciousness now, it's not like I really believed that it's like an idea you toy around with when you're trying to evade confronting the reality of your own mortality. Just the brain will produce all kinds of ridiculous forms of ignorance. And that was one I've been playing around with it. </s><s>[INST] You mean for like a large part of your life you were playing around with that. [/INST] I'm not like really? I think it's important to really emphasize, I didn't think I was immortal. Like I knew at some level I'm probably gonna die. Everyone dies, but there's ways that you can sort of poke around with that idea. I still do it to this day. Like I still do it. Like it's a natural thing to do when you're confronted with that with annihilation. You want a way out, you want to talk your way out, figure it out. There's gotta be some way to fix it. We'll fix it. That's another thing people do. They'll fix </s><s>[INST] it. They'll [/INST] expand the human lifespan. That's what they'll do. I mean that was, that's a big argument for it is like, look, the human lifespan up until covid, which they had to recalculate like the lifespan because of the statistically all the people who died, like threw it off a little bit. But pandemics aside the idea was the human lifespan seemed to be increasing by half a year every year or something like that. We're living longer. So all you gotta do one more half a </s><s>[INST] year and [/INST] we're immortal, right? And we can, we live a year longer every year, then we live forever. And so that's another way you can get out of confronting death as you can think. Well, maybe right now we don't have the tech, but it's coming consciousness uploads or downloads or whatever, depending on how you want to look at it. Another way people try to score them out of the reality of death. There's all kinds of tricks. </s><s>[INST] Yeah. And we do all of them. And sometimes, yeah, I mean, a lot of religions provide different, even more tools in the toolkit for coming up with ideas of how you can living the illusion that we're not going to. There's not an end to this particular experience that we're having here on earth right now. And then when you get that cancer diagnosis, it's like, yeah. What was that? Like going home the car ride, you drive home alone. [/INST] Yeah, I mean, it was one of the most. </s><s>[INST] Would you listen to Bruce Springsteen [/INST] Bruce Little Girl Is your daddy? That's </s><s>[INST] Not a Good 1 to listen. [/INST] You have cancer. Is he gonna die? Yeah. </s><s>[INST] Love songs? Maybe you take you experience them more intensely. [/INST] I don't remember what I listened to and I don't remember driving home, but I do remember driving to another doctor appointment. Doctor's appointment the next day. I think it was the next day. I think the Goodyear blimp was floating in the sky and I was looking, I was a stoplight, looking around is the, is the person flying and how to cure cancer? </s><s>[INST] No, [/INST] I didn't think that what I thought was, this ship just keeps going. That's what I thought. I thought I'm gonna be gone and it's just gonna keep going. And that was a beautiful moment for me. Was this beautiful moment </s><s>[INST] of like, [/INST] Oh yeah, you know, like, that's just what you're talking about with um the Ukraine, what you're talking about. It's like, unless you've been been there, it's really hard to explain to people that even in the midst of what is generally accepted as one of the worst things that can happen to you war cancer. Somehow, </s> |
<s>[INST] To not speak a certain way? No, like a little more civility? [/INST] I was just trying to get a second opinion on this. Second opinion of a- Non-internet people are way more extreme than- Yeah. Way more extreme. No, that's not true. Way more extreme. No, that's not true. Okay, so here's something for me, okay? I was not online until three years ago at all. I would watch YouTube. That's pretty much all I would do. I wouldn't do anything else. Really, I didn't grew up playing video games or anything. So I'm extremely new to everything. So when I came into this world, and I started seeing clips from him in the past, I didn't think I really had much of an opinion, because it just sounded like it was just a different words over mountain mountain mountain. I didn't mean anything. That's what it feels like if you're saying the r word, it's because you just want to call someone stupid. But you want to do it a little bit more. But it's not like, it didn't feel like it was like a, it's not like racist more. </s><s>[INST] Like, you know, it's not. You know, it's not agreement on this side, right. [/INST] So if he was saying the F word because it was just like a word to like insult someone and he was like, I don't think he was ever I don't think you're ever homophobic back in the day or anything like that, but I think it was just like a way to express yourself. Maybe back then, I don't know, I didn't do it. There's no videos of me or anything </s><s>[INST] because I wasn't even online back then. In my case was I definitely don't think Stephen is homophobic or is there any of those obviously? So there's a good heart there and a good mind. [/INST] I was just saying, it's not- </s><s>[INST] He just likes being mean. Well, there is some, you lose yourself and forget the bigger picture that he's pushing for more effective discourse on the internet. He's like an inspiration to a lot of people, especially now, of like how you can use effective conversation to make for a better world, to radicalize people and so on. And then you lose some of that power by losing yourself in like the language, just more language of emotion versus effective communication. [/INST] But it's a gray area, but- I would say like something that is probably recently done in that case, because he's been joking about women a lot, it's women's fault, they're bad. It's been like a lot of jokes when it comes to misogyny, I guess, in your community. And I think it's actually turned people a little bit that way. That's why we've done the recent misogyny, that you've been listening to a lot about. Yes, so, I guess that's actually true because I don't think it was pretty, I don't think it was clear enough. I don't think it actually was. I think you did that mistake but I think back then, I was even saying like, K, you should probably not, like you probably, you probably should not do that. Cause it actually is pretty hard for me because whenever I come into his community, like his chat, people are just gonna spam. It's like a woman moment. It's a woman moment whenever I say something and it's kind of like, yeah, it's getting pretty annoying. As I said, it's just annoying when you see it every single day. </s><s>[INST] And- There you go. [/INST] Wisdom from somebody younger than you. Wisdom can come from all kinds- Yes, just sometimes- Very limited quantities. </s><s>[INST] Learn something from anybody. What advice would you give to young people, the both of you, that you have both audiences where young people look up to you? In general, if you were to give advice to somebody in high school, like how to create [/INST] a life that can be proud of, what would you say? The most important thing that I've learned is to view people as different and not better or worse. And when you view people as different instead of better or worse, you learn that there's almost something that you can learn from anybody. Be open and empathetic towards other people's experiences. Nobody does anything by random choice. There's always reasons why people act the way they do. And as long as you're willing to be open and receptive to the lived experiences of other people, you're going to be able to gather information and create a more cohesive and </s><s>[INST] better view of the world than any of your peers will. [/INST] Do you have any kind of advice you can give to young folks? I feel like something that I see, especially in America a lot, is that a lot of people kind of get told what to do early on in high school, they're supposed to become this thing, education-wise, they're supposed to become a doctor or this thing or whatever, and then they kind of just give up on things that they're actually passionate about. I think a lot of teenagers get really confused, they get an education, and then they get that job and they hate everything. And they think that when they're reaching the job, when they're reaching the journey, they're going to get happy. That's where the happiness is going to be. But then when they get to there, they just hate everything and then become really depressed. And I've seen this so much. Like I see this all the time and it's pretty sad to me to see so many people that are just wasting time. And then they just get really confused and I don't know, it's the same thing with relationships too. No one really knows what they want anymore. I feel like everyone is just kind of doing whatever like society is saying, or the parents are saying, or their friends are saying. And they never really doing anything that's super meaningful anymore. And like they don't... So what I would say is like, try to find something that is important to you. It could be anything really, some sort of passion. Maybe your friends, maybe what matters to you? Figuring those things out, I think, is really important. </s> |
<s>[INST] What stores in [/INST] Thorazine is like a major anti psychotic drug. It's a it's a horrible drug but it knocks people out tranquilize them, we would never do that today. Um You know, we don't aboard a difficult experience like that, but in any case they hit that, that was not part of the Um write up of this experiment. So what they did is, in a sense, a little bit exaggerated the benefits. It later became a three years later after the experiment of four years in time magazine and said everybody that got still Simon had a mystical experience. Like it wasn't true. Not everybody ate out of attended but not all 10 not this guy. And they minimize the risks. So there was a bit of that. I think tim was reckless in that way, it was underplayed the risks and over promised the benefits and then the concord prison experiment. Um It turned out that uh tim had fudged the data completely and it wasn't really successful. So I fault him for that, that the outside world was doing the opposite. It was exaggerating the risks and blocking research. He felt justified to fight the data </s><s>[INST] because the outside world was fudging, in a sense, the response to the. Yeah. Yeah so so that that that presents very nice context fuck the government. But I'm glad that somebody uh is fighting the good fight from within and doing it the right way, which is Where you are. So the 80s, let me ask what is uh maps the Multidisciplinary Association for Psychedelic Studies and what what is its mission throughout the years. Throughout the decades. [/INST] Yeah. So uh maps is a nonprofit organization that created it as a nonprofit pharmaceutical company. Um I created an 86 after D. E. A. The Drug Enforcement Administration criminalized M. D. M. A. In 1985. And that was after they started trying to do that in 1984. And as I mentioned this uh terence McKenna sponsoring, you know motivating us to do this safety study. So so we did that in preparation for this eventual crackdown because M. D. M. A. was called Adam used as a therapy drug but it was also beginning to be sold as Ecstasy as a party drug and that was taking place in public settings and bars. And so it was inevitable that the crackdown would happen. And so I had a nonprofit connected to Buckminster Fuller earth metabolic design lab that we used to support this lawsuit against the D. A. To block them from criminalizing M. D. M. A. We were winning in the court of public opinion and winning in the court the D. A. Freaked out in the emergency scheduled M. D. M. A. N. 85. The handwriting was on the world that they were not going to permit the therapeutic use to continue because it gets in the way of the narrative of the drug war and these are terrible drugs. So in 86 is when I started maps as a nonprofit firma because the strategy that I realized is that americans are open to medicines, you know that tools to ease suffering. That was the opening wedge, the opening door to changing attitudes and it would be through science. I would say that my religion is more science than anything else and you know culture and religion are metaphorical but often too much they become literal but but I felt that through science through medicine um there would be way to bring these drugs back to the surface and the mission was always this mass mental health. This idea that what we need is to spiritualized humanity. Einstein said um the splitting of the atom has changed everything except our mode of thinking and hence we drift toward unparalleled catastrophe which shall be required if mankind is to survive as a whole new mode of thinking. So what is that new mode of thinking? I my presumption is that it's more of this mystical sense of thinking that that we're all connected and then if we realize that we're all connected we're not going to blow up the world. So a lot of people say that you know if we could just give LSD all the world leaders that would be you know then they have these spiritual experiences, the world would be better but actually had a ketamine experience the day after that D. M. T. Experience I described with the inner Hitler, this ketamine experience was I was above and behind Hitler as he was giving a speech in the Nuremberg rallies kind of thing and I was trying to think how do I get into his head, how do I undo what he wants to do, how can we deal with him. And and I realized this whole new thing about the Heil Hitler salute. And you know, he would like push energy out and then everybody would do the salute back to him. And so it's like the one of the many and the many to the one giving all these people, giving away their power and then how it would just sort of ratchet up in intensity like these vibrations. And I realized there's no way to get into his head. This idea we've talked about before about you have to be willing. So what that sort of helped me understand is that the strategy has to be mass mental health. It's not about changing a few leaders. We need to change the mass of humanity to this new mode of thinking, this new spiritual way. So maps was a nonprofit pharmaceutical company focused on psychedelics. Big pharma wasn't doing this work, government wasn't funding it. So the only source of funds I thought would be through non profit donations. And that's been true up until just a couple of years ago now that we have the rise of these for profits. But that's because we've cleared out the regulatory obstacles. We've got more scientific data about the benefits funded through philanthropy. We've changed public opinion and there's a lot less zeal for the drug war. So all of those things have changed. But at the time it was mass mental health was the goal two tracks. One was drug development. The other was drug policy reform so that it's not just available to people who have a clinical diagnosis, but people who our personal growth or you know, they should have access to it as well. I did not know at the time that no drug had ever been made into a medicine by a non profit. That was really good. I didn't know that. I might have been a little bit more daunted. Um and actually that didn't happen for 13 more years. It happened in 1999. Um and that was the abortion pill RED 46 that was approved in Europe, but a controversial nobody, no pharmaceutical company would take it. And it was John D. Rockefeller of the 3rd through the Population Council, with the major donor being Warren Buffett and the Rockefellers. Uh And the buffets and some of the Pritzkers were involved in funding this. So that was the first non profit. But but the maps was designed as um from the very beginning, not academic research into psychedelics, but drug development. And that's a fundamental distinction. And that's why I think we're years ahead now if everybody else in terms of making a psychedelic assisted therapy into a medicine, because our goal from the very beginning was not knowledge, not academic research, it was practical, It was drug development. How do we create new social structure is how do we create legal access to these things Now? In december of 2000 and 14, we created the maps public benefit corporation. So maps is a nonprofit but you know, in our 35 years we've raised about 100 and $10 million in donations. Um, what I didn't know when I started maps and it took me um quite a few years. I didn't even know this till about 89 years ago. Um, was that in 1980 for Ronald Reagan had signed a bill to create incentives for developing drugs that were off patent. So M. D. M. A. was invented by Merck in 1912. It's in the public domain. These incentives are called data Exclusivity, which means that if you make a drug into a medicine that does no patent protection, nobody can use your data for a period of time to market a generic and that will effectively be well, it's five years you do pediatric studies, you get six months extension and where and we are being required. If we succeed in adults to work with adolescents with PTSD, it blocks a generic competitor from applying for until that 5.5 years overtakes FDA at least six months to review. So more or less six years of data. Exclusivity. 10 years in Europe is did exclusivity. So the story then became to the donors that you're not gonna have to give us money forever because we can make money selling M. D. M. A. But we want to do to revolutionary things. You can say one is psychedelic assisted psychotherapy, but the other is marketing drugs when you market it with the profit maximization motive, we end up in the extreme getting the distortions that we have in America where we have the most expensive Health care system in the world per capita. But our outcomes are down like 40 or 50 among the countries are average outcomes. We don't have. Third of the people are, so don't have insurance and it's just very inequitable. So what we're trying to do is show a different way to market drugs. And it's a modification of capitalism is called the benefit corporation where you maximize public benefit, not profit, you still make a profit. So selling M. D. M. A. For a profit is not something we could keep inside the nonprofit because it's taxable, it's it's a business. So we've created the maps public benefit corporation which is 100% owned by the non profit. So we have a nonprofit that owns a pharma company. And </s> |
<s>[INST] to clarify the house on fire, meaning the cyber war that's going on, cyber attacks [/INST] in the 2016, you know, when the Russians had interfered in the election, you know, I mean, basically, this was a huge national security crisis and our politics, we've gone mad as a result of it, and in fact, we were making the situation worse and I felt that I could, you know, kind of at the time, but maybe I could do something here, I could try to clarify, I could, you know, work with others who I knew in the government from previous stints in the government to push back against this and try to make sure it didn't happen again. And look, and I also didn't have this, you know, mad, you know, kind of crazy ideological view of Russia either. I mean, I knew the place, I knew that people had been sitting a long time and quite calm about it. I don't take it personally, it's not kind of an extension of self it's, you know, something I've spent a long time trying to understand for myself going back to the very beginning of my with the Russians trying to blow us up. There must be an explanation. There was it was a very complicated and complex explanation. It wasn't as simple as how it sounded. And also there's a long tail to 2016, you know, Putin's perceptions, the kind of things that he thought were going on the, you know, the whole way that what they did was actually fairly straightforward. They've done this before in the soviet period during the Cold War classic influence operation. It just had gone beyond the bounds of anything they could have anticipated because of social media and just a confluence of circumstances, the United States as well. We were very fragile and vulnerable. And I remember at one point having a discussion with the Russian ambassador where, you know, we were complaining about the Russian intervention, he said, Are you telling me that the United States is a Banana Republic, that it's so vulnerable to these kinds of efforts? He actually looked genuinely mystified. Although he was probably part of a kind of political stick there. But he had a point, the United States had never been that vulnerable as it suddenly was in 2016. And in the time that I was in government and going back to what you asked about the whole impeachment and the whole exercise in in Congress. That vulnerability was as stark as it, you know, ever could be. Our domestic politics were as much a part of the problem as anything else. They were the kindling to all of the kind of the fires Putin, didn't start any of this other kind of problems domestically just took advantage of them. And you know, basically added a bit of an ex element here and there. </s><s>[INST] Yeah, the interference. I mean that's a much longer discussion because it's also for me, technically fascinating. Um, I've been playing with the idea of just launching like a million bots, but that are doing just positive stuff and just being kind Yeah, [/INST] it was kind of wondering if is it possible to do something on this skill that's positive? Because you know, a lot of people seem to be able to use all of this for pretty negative effect. You've got to kind of hope that you can do this. Use the same networks for positive effect. </s><s>[INST] I think that's actually where a lot of the war, I think from the original hackers to today, what gives people like me and I think a lot of people that in the hacking community pleasure is to do something difficult, break through the systems and uh, do the ethical thing. So do the, because if there's something broken about the system, you want to break through all the rules and do something that, you know, in your heart is the right thing to do. I mean, that's what Aaron Swartz did with releasing journals and publications that were behind pay walls to the public and got arrested for and then committed. Uh, but to me it's fascinating because I maybe you can actually educate me, but I felt that the Russian interference in terms of social engineering, in terms of bots, all that kind of stuff. I feel like there was more used for political bickering than to actually understand the national security problem, because uh, I would like to know the actual numbers involved in the influence. I would like to I mean, obviously, hopefully people now understand that better than are trying to defend the national security of this country. But it just, it felt like like for example, if I launched one bot and then and then just contact somebody at the new york times saying I launched this one bot, they'll just say uh, M I T scientist hacks, you [/INST] know, and </s><s>[INST] then that, but [/INST] that's exactly what happened. It was, you know, kind of, I think that, you know, Putin and some of the people around him understood because again, propaganda state, they spend an awful lot of time thinking about how you basically put out your own content and how you get maximum effect through performance. Putin himself is a you know, political performance artist. I mean, trump understood exactly the same thing, they were actually operating in parallel, not in collusion, but in parallel, you know, basically trump understood how to get lots of free air time, you know, how to get himself at the center of attention, Putin, you know, did that through a kind of, I think a less organic kind of way, you know, a lot of people working around him and that's the old bolshevik, agitprop and you know, kind of then the whole soviet propaganda machine and you know, Putin kind of growing up in that kind of environment and having, you know, the kind of the Kremlin press office and all the kind of people around him got kind of a massive machine, knew how that worked. I mean, they haven't done what the chinese did in Russia, like blocking everything and having a big firewall, it's kind of putting out lots of content getting into the, you know, the sort of center of attention trump's doing the same kind of thing. And the Russians understood that, you know, if you put a bit of things out there and then you call up the new york times and people are going to run with it and what they wanted was the perception that they had actually swayed the election. They loved it. This was the huge mistake of the democrats, everything. And I kept trying to push against this. No, they did not elect donald trump americans elected donald trump and you know, the Electoral college was a key part. Vladimir Putin didn't make that up, you know, and basically, I also remember, you know, one point, um Russian ambassador, you know, talking to me about when we were doing the standard, you know, here we are, we're lodging our complaint about the interference, you know, he didn't he basically said, well we didn't, you know, kind of invent Comey and you know, basically the, you know, the decision to reopen, you know, Hillary Clinton's emails or you know, kind of Anthony Weiner and you know, kind of his um you know, emails on his computer and I was like, yeah, he's right. I mean, you know, there were plenty of things in our own system that created chaos and tipped the election, not, you know, kind of what the Russians did, but you know, it's obviously easier to blame the Russians and blame yourself when, you know, things are kind of or those random forces and those random factors because people couldn't understand what had happened in 2016, there's no hanging chads like 2000 where there was, you know, kind of a technical problem that actually, you know, ended with the intervention, the Supreme Court, there was, you know, pure and simple the Electoral College at work And a candidate that nobody had expected, including the Republicans in the primaries, you know, to to end up getting kind of elected or put forward different 2016 suddenly becoming the president and they needed a meta explanation. It was much better to say Vladimir Putin Vladimir Putin, and the Kremlin guys were like, oh my God, yeah, fantastic Champagne corks popping this is great, our chaos agent, They knew they hadn't done it, but they'd love to take credit for it. And so, you know, the very fact that other people couldn't explain these complex dynamics to themselves basically dovetails beautifully with Vladimir Putin's attempts to be the kind of the Kremlin Gremlin in the system. And he's, you know, basically was taking advantage that forever more. And I wanted, you know, to basically try to work with us to cut through that. And the thing is then, you know, people lost faith in the integrity of the election system because people were out there, you know, suggesting that the Russians had actually distort the election books about that. They said, you know, that they hacked the system when, you know, they were trying to hack our minds. But again, we were the fertile sold for this. I mean, we know this from Russian history, the role of the bolsheviks, you know, the whole 19 twenties and 19 thirties with Stalin, the fellow travelers and the, you know, socialist, you know, international, I mean, the Russians and the Soviets have been, we're at this fears about kind of pulling, you know, kind of people along and into kind of a broader friend, but it didn't mean that they were influencing, you know, directly the politics of of of countries, you know, large, that plenty of interventions. It's just that we were somehow, it was, it was a confluence of events, a perfect storm. We were somehow exquisitely vulnerable because of things that we had done to ourselves. It was what americans were doing to themselves. That was the issue. </s> |
<s>[INST] So [/INST] as I state as I state these things, I know that I know I have protection if I have God's protection, that's why I'm here to this date, right? So it's for me to have this platform and express exactly what I feel because it's kids out there right that um it's kids that are gonna save the world through engineering and through facts and I've got to get download as much of the information and as much of the don't be afraid to state your facts is the biggest thing because the world is being ran by fear and that is no actually God runs the world, but there's just like little cloud this patina of our ego uh that deals with the money and the car and the girl were dating and all this in the clothes you wear and spending too much on clothes and a lot of stuff that I've been involved with promoting. So now what I'm promoting is you have the idea you say it out loud, like you like if you had Tourette's say your truth out loud, if you hate you hate me, say that out loud, say whatever you feel out loud, like you say it non violently, you know Nonviolent, nonviolent have to say that like as you know, you know as a shoutout to Alex jones, you know and trump where they try to say that when they say their truths out loud that it's inciting violence. So let's be really clear. I'm saying that they have criminalized free thought the you know they I hate when people use they like they get to blame for everything, right? It's always they who the who the funk is they? Exactly. The problem </s><s>[INST] is they is us. [/INST] Yeah. Well I love that what we needed to do is turn I two we so we take the responsibility and turn they to us. So we take accountability </s><s>[INST] to [/INST] us. They i to we and they to us and that relates it's a language at the same time you're seeing the language of our surroundings, You're seeing the language of one of things. Okay, let's go back to the sale explanation because I'm also doing it. What's that thing when Ellen just put all the information and open and people can figure out if they can figure it out and open source. </s><s>[INST] So I'm [/INST] open sourcing this idea right now so that you know, engineers and anointed people beings anointed beings can collectively contribute to this to push our species for it. Um So where I've got to with my research is that there will be a hole at the top that allows natural light, natural air and there's a constant water system. So it becomes like a water city where it's a constant flow, uh that's not, its its regenerative. It's not wasting the water and it's just a constant flow. So I put the toilet really close to where this uh this I don't want to call it a wave pool. But in the pool the water is not still so </s><s>[INST] that exists in water. Well, [/INST] the living room is that you have two rooms, you have the dry room which people call the living room before and then you have the wet room. </s><s>[INST] How is that a source of happiness? What what is broken about our world today? That that gives you in terms of the pursuit of happiness, which is one of the things [/INST] previous </s><s>[INST] declaration of all [/INST] of our buildings are based on just the industry and the economy. And then people lean into what people are used to seeing like what we think is attractive because then you go back to the, you know, women who give life so everyone wants to be attractive to the attractive. And that's the reason why the media goes and gets, you know, Bella Hadid kim Kardashian, Emily Ratajkowski. Um Those are the only ones I'm gonna uh </s><s>[INST] there's more, [/INST] there's more, but I'm I'm thinking of the ones that I think are </s><s>[INST] attractive. It is subjective [/INST] because it's subjective. So to my subjective, which I think I think uh the only guys that could argue with me on on my opinion on girls would be brad and leo because we're the only ones with the track record, you know, everything else. So for everybody else's subjective, we've got history. So are </s><s>[INST] you, are you are you happy? [/INST] Absolutely. I'm just here. It's actually, it's difficult to make me unhappy. Now. I can deal with frustrations, but when those frustrations are, it's like when you've allowed people to be in your life that shouldn't be there and then they do the thing you knew they were gonna do and you're screaming at them, but you really need to scream at yourself. </s><s>[INST] Yeah. So you realize anything that's like anger or frustration, That's something that's wrong with you. That needs to be dealt with. [/INST] Yes. Like right now, okay, say I got this pattern cutter that I wanted to remove because I felt like she was copying our patterns into the computer and I didn't know why she wanted to do that. And I just wanted to do the patterns by hand. And then, you know, </s><s>[INST] you mean like automating the patterns that me and [/INST] my head of it because they just that's the whole thing. It's like, it's a war, it's a war of human beings and robots, </s><s>[INST] but that makes it more efficient, right? You don't like the efficiency. You like the creativity of the singular nature of the pattern. I [/INST] believe that the human beings and I'm saying this to you as a person who makes robots and a person is friends with </s><s>[INST] deeply offended right [/INST] now. Yeah. </s><s>[INST] But yeah, I'll let it go. [/INST] Okay. So yeah, I believe that the </s><s>[INST] hates robots. I'm not gonna put that out. It's a joke [/INST] robots have feelings too. </s><s>[INST] Yeah, [/INST] that was also a joke </s> |
<s>[INST] and everybody. So the miners are incentivized, do not mess with the system because they're getting value from the system. So if they mess with it, it's going to decrease the value of their physical work investment. [/INST] Yeah, so they have to incur a real physical cost to produce a block. Right? So right now You get 6.25 bitcoins in a block at a minimum. And then maybe some fees as well. </s><s>[INST] How hard is it to produce a block now? [/INST] Well challenging I mean you need uh So 6.25 bitcoins and bitcoins worth $55,000 or so. So it's probably going to cost you about that amount to produce it because it's a free market competition and miners have very thin margins. So it's like if I auction off a dollar you would pay up to 99 cents to buy that dollar for me exactly what happens with minors there. You know, basically competing for the right to obtain new units of money. So logically speaking they would pay up to the value of that money in order to earn it. And </s><s>[INST] for people not familiar the process of mining is solving a difficult cryptographic problem. That's a computational problem. [/INST] It's I would say it's not like people sometimes representatives like a really challenging puzzle like the individual puzzles. Very simple like you can do with pen and paper if you wanted. You know like shot to 56. It's just that you're searching through the big mathematical space to find the needle in the haystack, you're just doing lots of iterations of a simple puzzle. </s><s>[INST] It's just brute force. Hence like the stability of the whole idea of the proof of work if it was if you if there was a shortcut it wouldn't be it wouldn't work. [/INST] Exactly. So let's hope nobody solves shot to 56. </s><s>[INST] Yeah there's a lot of discussions in from the quantum computing space but everybody I I talked to all my colleagues and that work on quantum computers uh say that we're quite a way quite a long way away from that being an issue in cryptography and then and certainly an issue currency [/INST] that should have been one of the sides on these dice and should have been quantum quantum because I don't think it is. I forgot to put on this edition. </s><s>[INST] People should check out scott Anderson uh there's a lot of people that are kind of uh selling quantum snake oil so you should be very careful. I think it is really exciting space that might change the world in the next decade or 100 couple 100 years especially for simulating quantum mechanical systems. But in quantum machine learning people should check out tensorflow quantum. It's a nice way to sort of educate yourself about the space. And actually if you're pragmatically minded too uh you know through software engineering, explore how you simulate quantum circuits, how you run machine learning on those quantum circuits. The main point that scott makes scott Aaronson people should check out his blog too is that like there's not yet a single machine learning application that doesn't do almost as well in a classical computer. So it doesn't like yes the dream is somehow quantum computers will change the nature of artificial intelligence but there's yet to be an actual algorithm that uh or problem set or data set where that would be the case. So uh Skepticism is good in this space anyway. That says you kind of explained how Bitcoin works. You also wrote a blog post recently giving a shout out to the new book. The block size wars. What is a block size? What are the block size wars? Its history, its importance. It's a philosophical foundations. [/INST] Yeah, I mean Bitcoin at this point we have our own civil wars. If you're wondering about how politically intense it gets, </s><s>[INST] it's currently not hot, it's cold. It's [/INST] oh yeah, we're in a detente right </s><s>[INST] now. There's no tanks or missiles at least not yet. Hopefully [/INST] it can get a little violent I guess. I think one of the Bitcoin core developers or one of the participants in the war got swatted at one point. What swatted means when someone does a fake phone call saying that you're holding someone hostage at your house and the swat team goes </s><s>[INST] pretty school [/INST] internet warfare tactic. Yeah, but the block size war I would say effectively ended. Although we're definitely going to have more civil wars in Bitcoin for sure. But basically the core argument was a technical one on its surface but a very deep political one at its core. The technical question is how many megabytes should be in each successive block. So Satoshi basically installed a limit of one MB per block. </s><s>[INST] So we should backtrack there was no limit in the beginning and it seems like Satoshi, what is this? 2000 The war started in what 2017 or something like that? I don't know in [/INST] The 15 was when the, you know the battle cry and began, what was the first battle of the civil war? Remember? </s><s>[INST] But sort of Satoshi, I don't know if you can comment at on it. Like why does until she set the limit to one megabyte all of a sudden almost secretively and in the beginning there was no limit whatsoever. [/INST] Yeah, I mean we can get into and people have spent thousands of hours poring over Satoshi is writings to find, you know which side Satoshi was a line and you can find like any textual exegesis, you can find evidence for either side. Right? But yeah, I mean effectively when Bitcoin was launched, there was a block size because if you made a block over a certain size with the first edition of the code, it would have crashed nodes. But then in 2010, Satoshi added the one megabyte limit in a covert way with no comments or anything. And that stuck basically and uh then Bitcoin blocks filled up and people that have been socialized into this vision of Bitcoin as an effectively free transactional network, like why pay transaction fee if you're not at congestion, if the block isn't full the minor will mind your transaction for free. Right. Um people that had been brought up in that status quo From 2009 to kind of 2015, they noticed the black started to fill up and they're like, okay, well let's just remove this arbitrary limit, right? What could possibly be the harm? And then a whole other factions said, no, you need to cap the data throughput of the system because if you increase it, it's going to be highly exclusionary. Uh and ultimately, regular folks are not gonna be able to run a full node. </s> |
<s>[INST] How much anxiety or even fear do you have before making a difficult decision on the chess board? So it's a high stakes game. How nervous do you get? How much anxiety do you have in all that calculations you're sitting there for 10, 15 minutes because you're in a fog, there's always a possibility of a blunder of a mistake. Are you anxious about it? Are you afraid of? [/INST] It really depends. Um I have been, I have been at times. I think the most nervous I've ever been was Game 10 of uh the world championships in 2018. Although that was just a thrilling game. I was black. Um I basically abandoned the queen's side at some point to attack him on the king's side. And I knew that my attack, if it doesn't work, I'm gonna lose. But I had so much adrenaline. So that was, that was fine. I thought I was gonna win. Then at some point I realized that it's not so clear and my time was ticking and I was just getting so nervous. I I still remember what happened like we played this time. Trouble face where he had fair little time, but I had even less. And I just remember, I cannot remember much of it just that when it was over, I was just so relieved because then it was clear that position was probably gonna peter out in um in a draw. Otherwise I'm often nervous before games. But when I get there, it's all business and especially when I'm playing, well, I'm never afraid of of losing when I when I play because I trust I trust my instincts, I trust my, my skills, </s><s>[INST] how much psychological intimidation is there from you to the other person, from the other person to you. [/INST] I think people would play a lot better if they played against an anonymous me. I would love to </s><s>[INST] you. [/INST] I would love to have a tournament online where let's say you play 10 of the best players in the world and you don't for each round, you don't know who you're playing. </s><s>[INST] That's an interesting question. You know like there's these like videos where people eat Mcdonald's or burger king or diet coke versus Pepsi. Would people be able to tell they're playing you like from the style of play do you think or from the strength of play [/INST] if there was a decent sample size? Sure. What </s><s>[INST] about you? Would you be able to tell others [/INST] into in just one game? Very unlikely. What </s><s>[INST] what sample size would you need to tell accurately? I feel like it's a science. [/INST] Yeah, I think 20 games would help a </s><s>[INST] lot. [/INST] Yeah. But I know that they've already developed ai butts that are pretty good at recognizing somebody's style </s><s>[INST] which [/INST] is um which is quite fascinating. Uh </s><s>[INST] And they'd be fascinating if those bots were able to summarize the style somehow. Maybe great attacking chest. Like some of the same characteristics you've been describing, Like great short line calculations, all that kind of [/INST] stuff. </s><s>[INST] I mean really, [/INST] really all the best chess players, there are basically just two camps? People who are good at long longer lines or or shorter lines. It's the hare and the tortoise basically. And sometimes you know, I feel like I'm the closest you can get to a hybrid of those because </s><s>[INST] you've got both your good in every position to the middle game and end game. [/INST] Yeah. And also I can I can think to some extent both rapidly and deeply which a lot of people that can do both. But I mean to to answer your question from before I think yeah, I sometimes can get a little bit intimidated by my opponent but it's mostly if there's something unknown, if it's mostly if it if it's something that I I don't understand fully and I do think especially when I'm playing well people, they just play more timidly against me than they do against each other sometimes without even realizing it. And I certainly use that to my my advantage. If I sense that my opponent is apprehensive, I sense that they're not gonna necessarily take all their chances. It just means that I can take more risk and I always try and try and find that balance </s><s>[INST] to shake them up a little bit. Yeah. What's been the toughest loss of your career that you remember? Would that be the the World Championship match? [/INST] Oh yeah, sure. </s><s>[INST] Can you take [/INST] 2016. </s><s>[INST] And who was it against [/INST] against in new york? You </s><s>[INST] Take it through the story of that game, where were you before that game in terms of game one through 7. [/INST] Yeah, so game one and two and not much happened. Games three and four, I was winning in both of them and normally I should definitely convert it both. I couldn't partly due to good defense on on his part, but mostly because I just I messed up. Uh and then after that games five, six and seven, not much happened, I was getting impatient that at that point. Um so for Game eight I was probably ready to take a little bit more risks than I had before, which I guess it was insane because I knew that he wouldn't beat me unless I beat myself like he wasn't strong enough to play me and </s><s>[INST] that was leading to impatient somehow and impatience. No, [/INST] Because I knew that I was better. I knew that I was better. I knew that I just needed to win one game and then the match is over. That's what happened in 2021 as well. Like when I won the first game against, I knew that the match was over. Unless I like funk up Royally then he's not gonna be able to beat me. So what happened was that I played a kind of an innocuous opening as whites just trying to get a game trying to get him out of book as soon as possible. Then </s> |
<s>[INST] you're fully conscious, there's something going out there. It could be just a side effect. So what is the the [/INST] you mean the phenomenon? </s><s>[INST] So what's the selective meaning? Why? Uh what what is the function of you being able to lay in this uh sensory free deprivation tank and still have a conscious experience [/INST] revolutionary. Obviously we didn't evolve with flotation tanks in and our environment. I mean, so biology is notoriously bad at asking why? Question to economical question, why do we have two eyes? Why don't we have four or three years or something? Well no, there's probably there's a function to that. But we're not very good at answering those questions. We can speculate and leslie where biology is very or scientists very good about mechanistic question why that charge in the universe, right? We find ourselves in a universe where the positive negative charges, Why why does Quantum mechanics Hall you know why why why doesn't some other theory, whole quantum mechanics or in our university is very unclear why. So tell economical question why questions are difficult to answer clearly there's some relationship between complexity, brain processing power and consciousness. But but however in these cases in these three examples. Okay I gave one is an everyday experience at night. The other one is a tom and third one is in principle you can everybody can have these sort of mystical experiences. You have a dissociation of function from of intelligence. From uh from consciousness </s><s>[INST] consciousness. You called me asking a why question. Uh let me ask a question. That's not a why question. You're giving a talk later today on the turing test for intelligence and conscience of drawing lines between The two. So is there a scientific way to say there's consciousness present in this entity or not? And uh to anticipate your answer because you you also there's a neuro biological answer. So we can test the human brain. But if you take a machine brain that you don't know tests for yet how would you even begin to approach a test if there's consciousness present in this thing? [/INST] Okay that's a really good question. So let me take in two steps. So as you point out for for for for humans let's just stick with humans. There's now a test called zap and zip is a procedure where you ping the brain using transcranial magnetic stimulation. You look at the electrical reverberations essentially using E. G. And then you can measure the complexity of this brain response and you can do this in a wake people in a sleep, normal people. You can do it In a wake people and then anesthetized them. You can do it in patients and it it has 100% accuracy that in all those cases when you're clear the patient or the person is either conscious or unconscious, the complexities either high or low. And then you can adopt these techniques to similar creatures like monkeys and dogs and and and mice that have very similar brains. Now, of course, you you point out that may not help you because we don't have a cortex, you know, and if I send a magnetic pulse into my iphone or my computer, it's probably going to break something. So we don't have that. So what we need, ultimately we need a theory of consciousness. We can't just rely on our intuition. Our intuition is. Well, yeah, if somebody talks, they're conscious. However, then there are all these pages Children babies don't talk right, But we believe that the babies also have conscious experiences, right? And then there are these patients I mentioned, uh, they don't talk when you dream. You can't talk because you're paralyzed. So, so what, what we ultimately need. We can't just rely on our intuition. We need a theory of conscience that tells us, what is it about a piece of matter? What is it about? A piece of highly excitable matter like the brain or like a computer that gives rise to conscious experience. We all believe none of us believe anymore in the old story. It's a soul that used to be the most common explanation that most people accept that. It's still a lot of people today believe, well there's a, there's God endowed only asked with a special thing that animals don't have renamed the Carp famously said dog, if you hit it with your carriage may yelp me cry, but it doesn't have this special thing, It doesn't have the magic, the magic sauce. It doesn't have gas confidence. That's all. Now we believe that isn't the case anymore. So what is the difference between brains and and these guys Silicon? And in particular once a behavior matches. So if you have Siri Alexa and 20 years from now that she can talk just as good as any possible human, what counts? Do you have to say? She's not conscious in particular. If she says it's of course he will. Well of course I'm conscious you are, how you doing? And she'll say, well, you know, they'll generate some way to of course she'll behave like a like a person. Now there are several differences. One is so this relates to the problem. The very hard why is consciousness a hard problem? It's because it's subjective, right? Only I have it for only I know I have direct experience of my own consciousness. I don't have experience your consciousness now I assume as a sort of evasion, personal beliefs and probability theory and all of that. You know, I can do, I can do an abduction to the, to the best available facts. I deduce your brain is very similar to mine. If I put you in a scan of your brain is graphic, gonna behave the same way as I do. If if if you know, if I gave you this music and ask you, how does it taste? You tell me things that you know that that I would also say more or less so and for based on all of that, that you're conscious. Now with syria, I can't do that. So there I really need a theory that tells me, what is it about? About any system, this or this that makes it conscious. We have such a </s> |
<s>[INST] conscious self aware exactly to us it does make me a little bit sad as a human and just watching all the breakthroughs on artificial intelligence side when applied to natural sciences now more more to physics that the creatures that will solve the question of the origin of the nature of life or just the process the nature of life will be A. I. systems makes me a bit sad too [/INST] I don't think so </s><s>[INST] because you think that humans will [/INST] at this point in time remember was behind a I you know I I'm not buying in the singularity thing yet two A. I. is not aware A. I. is being built by humans so A. I. is a tool and extremely smart tool as long as we build it and as long as we use it as a tool it remains a tool and I think there is a lot of brew ha ha and of course the science fiction and movies they don't help </s><s>[INST] check I got to push back a little bit yes I agree with you for the most part in terms of ha ha ends scifi but there is I can work of DeepMind we can look at Chesil economic approaching folding so chess is a simple one to first look at what are Alfa zero which is their game playing engine was able to discover Nostoc fish about chess humbles the best human players not just is better than them it comes up with ideas that the humans don't understand and so the A. I. knows telling you even those programmed by humans the allies saying like sacrificing a upon here is a good idea sacrificing a queen bishop here's a good idea and then you start to kind of into it as a human why we don't deeply understand and you can say that a I was not conscious it doesn't deeply understand what humans do but there still are a wisdom and a depth of knowledge in that chess playing program that humans don't have in the same with the alfalfa folded protein folding there's a another planet to physics to sis simulating nuclear reactions and so on it feels like there might be a way to understand the nature of life that we can kind of into it poetically as humans but the the true understanding will come from a system that's much welcome competition sophisticated [/INST] again you know I would push back on on my turn because I still think that human give themselves the ability to do that by building that tool so the idea that the tool you know </s><s>[INST] what what what [/INST] we are getting into the car she scale and and and and the dark forest and all these things we can see the world this way I at this point in time for me I still see the great tool now are whether this I. F. I see are you always going to you know happened it's a try I still think that we are far away from from this but if that tool is capable of giving me a new perspective it's just that we're starting to jump into a deeper condition of of what the universe is whether it's through our brain though through a different way of gathering information remember this is what we do </s><s>[INST] yeah humans able to actually build tools and then like integrate them enter their way of thinking maybe another generation has to be born that is raised with those tools but we seem to like take for granted all the cool technologies you're integrate into your way of thinking a lot of people are growing up now their mind is integrated with the internet and yet you basically reconfigure the way you memorize things you no longer have to memorize a lot of fast casual look come up really quickly [/INST] yeah </s><s>[INST] and so like S. U. re allocate a lot of resources for thinking versus memory of just strict facts that kind of stuff we integrate all of that [/INST] yeah and and you know they are I would completely agree with you in fact I wrote about this again in in this new book that's coming out </s><s>[INST] what is the book coming up [/INST] in January it's evil been French actually just off with </s><s>[INST] you wrote in French [/INST] actually a rooted first in English </s><s>[INST] yes [/INST] and I translated into French so the the English version is already pretty much ready to go if we find a publisher in yet in the US but anyways the point of being here is that I looked at this as our relationship with technology as a complete change to me this is the singularity more than anything else which she is the co evolution of human with technology not anymore with their environment why we are missing of the environment right now why we don't respond to when they make the way we should because we are disconnected to the environment we are taking our information from and we were adapting from right now exactly as you said we take the information from the web from the phones it's a truck we have no filter over that information before you were out in the environment the information you get is the one the planets is sending you now this information is coming from different way you have no way of knowing if for missions correct or not you're gonna push back on that now yeah you look as this is an echo system and it explains a lot about behavior </s> |
<s>[INST] He was part of the group with Geoff Hinton from [/INST] which emerged the </s><s>[INST] back propagation paper. Anyway, I mentioned all that because I had a conversation with him about psychiatry. He also wanted to be a psychiatrist growing up as I have. And so very much believes in the magic of talk therapy, of exploring the human mind through talking. And so I think better help is worth trying. It's easy, private, affordable, available worldwide. Check them out at better help dot com slash lex. That's better help dot com slash lex. This is a lex Friedman podcast. And here is my conversation with Donald Knuth. Your first large scale program. You wrote it in IBM 6 50 assembler in the summer of 1957. I [/INST] wrote it in decimal machine language. I didn't know about assembler until a year later, </s><s>[INST] But the year 1957 the year. And the program is, [/INST] Yeah, I might have learned about assembler later that summer. I probably did 957, hardly anybody had heard of assemblers. You looked at the user manuals, how did you write a program for this machine? It was, It would say uh, you know, you would say 69, which meant load the distributor and then you would give the address of the number you wanted to load into the distributor. Uh, yesterday, my friend at Doug Spicer at the Computer history Museum sent me a link to something that just went on youtube. It was the IBM s uh progress report from 1956 which is, you know, very contemporary with 1957. Um, and in 1956 IBM had donated stanford University An IBM 651 of the first ones when they showed a picture of the assembly line for IBM six fifties and they said, you know this is number 500 or something coming off the assembly line. And I had never seen so many IBM 6 56. I did in this movie that was on Youtube now. Uh and and it showed the picture from stanford uh that you know, they said look, you know, we we donated one of these two stanford one to M. I. T. And they mentioned one other, one other college. And in December of 56 they donated to my university case tech. Um but anyway, They showed a picture then a class session uh where a guy was teaching programming. And on the blackboard it said 69 8000. I mean he it was he was teaching them how to write uh code for the IBM 6 50 which was in decimal numbers. So the instructions were 10-10 decimal digits. You had two digits and said, what what to do, four digits to say, uh what to do it too. And four more digits to say where to get your next instruction. </s><s>[INST] And there's a manual that describes what each of the numbers mean. [/INST] And Emmanuel was actually one. Uh If the manual had been well written, I probably never would have gone into computer science, but it was so badly written, I figured that I must have a talent for it because I'm only a freshman and I and I could write a better manual. Uh You did, and so I I started working at the computer center uh and And uh wrote some manuals then, but but but but this was uh but this is the way we did it. And and my first program then was June of 1957. </s><s>[INST] The tic tac toe. [/INST] No, that was the second program. The 1st 3rd program. The first program was factor in a number. So you dial the number on the on the uh switches. I mean you sat at this big man mainframe and you turn the dials set a number and then it would punch out the factors of that number on cars. </s><s>[INST] So that's the input, is the [/INST] input was yeah, the input, what, what was the number attended a number and and uh and the output was its factors and and and I wrote that program, uh I still have a copy of it somewhere. And um </s><s>[INST] how many lines of code do you remember? Well, [/INST] yeah, it started out as about 20, but then I kept hearing me debug it. I discovered debugging of course when I wrote my first program. And </s><s>[INST] what does the debugging look like on a program with just all numbers? [/INST] Well you sit there and you, I don't remember how I got it into the machine, but but I think there was a way to punch, punch it on cards. So each each instruction would would be 11 card. Or maybe I could get seven instruction on the card and instructions I don't know. But anyway so I'm sitting there at the at the council of the machine. I mean I'm doing this at night when nobody else is around of course um and and so you have one set of switches where you dial the number I'm inputting but there's another switch that you know that says okay now execute one instruction and show me what you do what you did. Or you or you there was another four switches and say stop if you get to those if you get to that instruction so I can see now go until you get there again and watch. Okay so I could watch you know it would take that number and divide it by two and if it's you know there's no remainder then okay to is a factor. So so then I work on but if if if not divisible by two, divide by three, okay keep trying and uh until you know you're you're at the end </s><s>[INST] and uh you would find a bug if uh if you were just surprised that something weird happened. [/INST] Well certainly I mean first of all I might have You know try to divide by one instead of two off by one errors. People make all the time but maybe I go to the wrong instruction, maybe I you know, maybe I I left left something in a register that I shouldn't have done, but the first bugs were pretty, you know, I probably uh on the first night I was able to I was able to get the factors of 30, you know, is equal to 23 and five. Okay. Um </s><s>[INST] so you're sorry to interrupt, You were so you're sitting there late at night, so it feels like you spent many years late at night working on a computer. So like what what's that like? So most of the world is sleeping and you have to be there at night because that's when you get access to the computer [/INST] between my freshmen sophomore year, I didn't need sleep. I used to do all night when I was in high school, I used to I used to do the whole student newspaper every monday night. I would, you know, I just stay up all night and and it would be done on Tuesday morning. Um uh that was, I didn't get ulcers and stuff like that until later, but uh </s><s>[INST] well uh I don't know if you know Rodney brooks, Rod brooks of course, yeah, he told, he told me a story that he really your, you know, he really looked up to you, he was actually afraid of you uh because he was [/INST] uh but </s><s>[INST] it when he tells the story when you were working on tech that they screwed up something with a machine, I think this might have been m I T I don't know. And you were waiting for them to fix the machine so you can get back to work late at night. Oh, that happened all the time. He was really intimidating. He's like doc and youth is not happy with. [/INST] That's interesting. But no, no, the machine at stanford Ai lab was uh was down an awful lot because we they had uh they had many talented programmers changing the operating system every day. And so operating system was getting better every day, but it was also crashing. So, so so I wrote almost the entire manual for tech during downtime. Uh but that's another story. </s><s>[INST] Okay, well, he was saying they it's a hardware problem, they they tried to fix it and they re inserted something and smoke was everywhere. He was, [/INST] it didn't happen as often as the operation. But yeah, it </s><s>[INST] was, it's a funny story because you're saying there's this tall, uh don knew that I look up to and I see it was pressure to, I think the computer. Well, it's funny, okay, the kind of things we remember that. Stick in your memory. Well, okay, [/INST] Yeah, well, I can tell you a bunch of Rod Brooks story. Stupid. Let's let's go back to the 50. So, Uh so I'm debugging this, my first program and I I had more bugs in it than a number of lines of code. I mean, the number of lines of code kept growing and let me explain. So I had to punch the answers on cards. Alright, so suppose I suppose I'm factoring the number 30. Then I got then I I gotta I gotta put too somewhere on the car and I got to put a three somewhere on the card. I got to put a five somewhere on the card. Right? And and you know what? It was my first program. I I probably screwed up and, you know, it fell off the edge of the card or something like that. But, but I didn't realize that there was some tentative numbers that have That have more than eight factors. Um, and the card has only 80 columns. And so I need 10 columns for every factor. So, my first program didn't take account for the fact that I would have to punch more than one card. My first program, you know, just lying the stuff up in memory and then I punched the card. But after, you know, so by the time I finished, I had to I had to deal with lots and lots of things. Also, I uh if you put a large prime number in there, my program might have said there for 10 minutes or 6 50 was pretty slow. And so I would sit there spinning its wheels. And you wouldn't know if it was in a loop or </s><s>[INST] whatever that tended is. Yeah, [/INST] So I think the largest, his sort of (999) 999 9997 or something like that. That would you know that that would take me a while uh For that first point anyway, that was my first program. </s><s>[INST] Well what was your goal with that program? You were hoping to find a large prime maybe? Or no, the opposite. You know, [/INST] my goal was to see the lights flashing and understand how this magical machine would be able to do something that took so long by hand. </s><s>[INST] So what was your second program? My [/INST] second program was a converted number from from binary to decimal or something like that. It was much, much simpler. Didn't have that many bugs in it. My third program was tic tac toe </s><s>[INST] and he had some, So the tic tac toe program, it's interesting on many levels, but one of them is that it had some, you can call machine learning in it. [/INST] That's Yeah, that's right. Uh I don't know how long it's gonna be before the name of our field has changed from computer science to machine learning. But uh but anyway, uh it was my first experience with machine learning. Ok, so here we had, </s><s>[INST] how does the program? Well, first of all, what is the problem you are solving? What is tic tac toe, What are we talking about? And then um Right, what, how was it designed? [/INST] Right, so you got you got a three by three grid and each, it could be in three states. It can be empty or it can have an excellent Oh alright so three to the ninth is a uh well what is, how big is it should know but it's 80 81 times 81 times three. So um Anyway eight is like two to the third and so that would be uh that would be like two to the sixth. uh but that would be 64 then you have to anyway </s><s>[INST] I love how you're doing the calculation. So a lot [/INST] of anyway </s><s>[INST] the three comes from the fact that it's either empty an X. Or an O. [/INST] And The 6:50. What was it was a machine that had only ah 2000. Yeah yeah 10 digit words you go from 0000 to 1999 and that's it and each word you have a 10 digit number so that's not many bits. I mean I got to have in order to have a memory of every position I've seen I need 3 to the 9th bit. Uh Okay but it was a decimal machine to, it didn't have this but it did have it did have strange instruction where if if you had a 10 digit number that but all the digits were either eight or nine. Uh You'll be 8998 or something like that. Would uh You can make a test whether it was eight or nine. That was one of the strange things IBM engineers put into the machine. I have no idea why. Well um hardly ever used but anyway I needed one digit for every every position I've seen. Zero met it was a bad position. I meant it was good position. I think I started out at five or six you know but if you if you win a game then you uh then you increase the value of that position for you but you decrease it for your opponent. Uh But but I could I had that much total memory for every every possible position was one digit and I had a total of 20,000 digits which had which had to also include my program and all the logic and everything including how to how to ask the user what the moves are and things like this. Okay so so I think I had to work it out. Get every every position in tic tac toe. It is equivalent to roughly eight others because you you you can rotate the board um which gives you a factor of four when you can also flip it over. And that's another factor too. So I might you know, so I might have needed only three to the 9th over eight positions of you know plus plus a little bit. Uh So I had but anyway that was that was a part of the program to squeeze it into this tiny, </s><s>[INST] so you tried to find an efficient representation that took account for that kind of rotate. I had to [/INST] otherwise I couldn't do the learning. Um So, but when I had three parts of my tic tac toe program Uh and I call it brain one brain to in brain three. So brain one just played a uh let's see it at random. Okay, it's your turn. Okay, You gotta put next somewhere has to go on an empty space. But that's that's it. Okay, choose Choose one and play their brain to uh had a candy routine. And I think it was, it also maybe it had maybe to assume you were the first player or maybe it allowed you to be first. I think you're hard to be either first or second but had a canned built in strategy known to be optimum for detect. Oh before I forget, by the way I learned uh many years later that Charles Babbage had had planned to have thought about programming tic tac toe for his, for his dream machine that he, that he was never able to finish. </s><s>[INST] So that was the program he thought about more than [/INST] 100 years ago. He had uh, he did that okay. And I had, and I had been influenced by a demonstration at the Museum of Science and industry in Chicago. It's like it's like boston's science museum, I think bell labs had had prepared a special exhibit about a telephones and relay technology and they had a tic tac toe playing uh machine as part of that exhibit. So that had been one of my uh, you know, something I had seen before. I was a freshman in college. And and inspired me to see if I could write a program for, okay, so anyway, I brain one random uh, knowing nothing brain to knowing everything. Then brain three was the learning one and I could, I could play brain one against brain one, brain one against brain to and so on. And so uh, You can also play against the user against the live in person. But uh, so I started going the learning thing and I said, Okay, you know, take two random random people just playing uh, detect oh uh, knowing nothing. Um, and after about, I forget the number now, but it converged after about 600 games uh, to a safe draw. The way my program learned was actually, it learned how not to make mistakes because uh it didn't try to do anything for winning. It just tried to say news. </s><s>[INST] So that [/INST] was probably because of the way I designed to learn anything, I could have had a different uh, reinforcement function that would, that would reward brilliant play. But anyway, it didn't and uh, and if if I took a novice against uh, you know, the skilled player, uh it was able to learn how to play a good game. So that was, and that was really my, but after I finished that, I felt I understood programming </s><s>[INST] was there. Did you did a curiosity and interest in learning systems persist for you? So Why, Why did you want brain 3 to learn? [/INST] Yeah, I think naturally it's, we're talking about Rod brooks. Uh he he was teaching all kinds of very small devices uh to to learn stuff um if a leaf drops off of a tree, uh uh you know, he was saying something while it learns if there's wind or not, but I mean he he pushed out a little bit too far, but he said he could probably train some little mini bugs to scour out dishes if he had enough a financial support. I don't like that. </s><s>[INST] Can I, can I ask you about that? Because um he also mentioned that during those years, there was discussion about inspired by touring about computation, you know, of what is computation? Yeah. And [/INST] yeah, I never thought about any stuff like that. That was, that was way too philosophical. I was I was a freshman after. I mean I didn't, I was pretty much a </s><s>[INST] machine. So it's almost like, yeah, I got you. It's a tinkering mindset. Uh not a philosophical mindset, [/INST] it was just exciting to me to uh to be able to control something, but not, but not, but not to say my solving a big problem or something like that? or is this a step for humankind? Right. No, no. </s><s>[INST] When did you first start thinking about computation in the big sense, you know like the universal turing machine? Well, [/INST] I mean I had to pass and I had to take, I had to take classes on compute ability when I was a senior. So you know, we read this book by martin Davis and yeah, this is cool stuff, but you know, I learned about it because I know I need to pass the exams but I didn't, I didn't invent any of that stuff, but I had create fun playing with the machine. You know, I um I wrote program because it was fun to write programs and get this. I mean it was like watching miracles happen. </s><s>[INST] You mentioned in an interview that when reading a program, you can tell when the author of the program changed [/INST] uh </s><s>[INST] how the heck can you do that? Like what makes a distinct style for program or do you think, you know there's different Hemingway has a style of writing versus James Joyce or something? Well [/INST] those are pretty, yeah, those are pretty easy to imitate, but it's the same with music and whatever you can. I found uh well during the pandemic I spend a lot more time playing the piano and I found something that I'd had, I had when I was taking lessons uh before I was a teenager and uh it was yankee doodle uh played in the style of, you know, and you had, you had, you had Beethoven and you had WC and Chopin and you know, and the last one was Gershwin and and I played over and over again, I thought it was so brilliant, but but it was so easy, but also to appreciate how this uh this uh author Mario somebody or other had have been able to reverse engineers the styles of those computers. So, but now particularly uh your question, I mean there would be there, it was, it was pretty obvious in this program I was reading, it was it was a compiler uh and it had been written by a team at, at Carnegie Mellon and I have no idea which program was responsible for, but you would get to a part where the guy would just not know how to how to move things between registers very efficiently and so and so everything that, that could be done and one instruction would take three or something like that, that would be a pretty obvious change in style, but there were, but then they were also flashes of brilliance where you could do in one instruction. Normally I used to, because you knew enough about the way the machine work that you could, that you could accomplish two goals in one step. So it was mostly the brilliance of the concept more than the semi colons and or the, you know, the use of short sentences versus long sentences. </s><s>[INST] So you would see the idea in the code and yeah, the the different style thinking. It was, yeah, [/INST] it was stylistic. I mean I could identify authors by their, by the amount of technical aptitude they had, but not by style in the sense of a rhythm or something like </s><s>[INST] that. So if you think about Mozart, Beethoven bach, if somebody looked at don canoe the code, would they be able to tell that this is a distinct style thinking going on here? What do you think? [/INST] Uh, </s><s>[INST] and what, what would be the defining characteristic of the style? [/INST] Well, my code now is, is literate programming. So I'm, it's a combination of english and see mostly, but if you just looked at the sea part of it, you would also probably notice that I don't, you know, that I use a lot of global variables that other people don't. And I expand things in line more than instead of calling anyway, I have different subset of see that I use. Okay, </s><s>[INST] But that's a little bit stylistic. Uh, but [/INST] but with literate programming, you alternate between english and, and see or whatever and um, </s><s>[INST] and by the way, people listening to this should look up literate programming. It's very interesting concept that you the proposed and developed over the years. [/INST] Yeah. Yeah, I'm that's the most uh, significant thing I think to come out of the tech project. It is that I I realized that uh programs were to be read by people and not just by computers and and that typography could massively enhance that. And and so uh I mean it they're just wonderful if they're going to look it up that they should also look up this book by called physically based rendering by matt Far and gosh anyway, it got an academy Award. But it's but all the all the graphic effects you see in movies uh you know, are accomplished by algorithms in this book is the whole book is illiterate program. It tells you not only how you do all the shading and and uh bringing images in that you need for animation and textures and so on. But it also you can run the code. Uh and and so I find it uh uh extension of the way I uh of how to teach programming. It is but by by telling a story as part of the program. </s><s>[INST] So it's uh it works as a program but it's also readable by humans. [/INST] Yes. And especially by me. Uh Yeah, a week later, a year later. </s><s>[INST] That's a good test. If you yourself understand the code easily a week or 30 year later. [/INST] So it's uh this place, it's the greatest thing since sliced bread programming or literate literate program. Okay. </s><s>[INST] Uh you heard it here first. Okay. You dodged this question in an interview I listen to. So when you ask you again here, uh what makes for a beautiful program? [/INST] What makes for a beautiful program? </s><s>[INST] What are the characteristics you see? Like you just said literate programming? What are the characteristics you see in a program that make you sit back and say, that's pretty good. [/INST] Well, the reason I didn't answer is because there are, there are dozens and dozens of answers to that because each you can define beauty the same personal. Do you find beauty in a different way from our our I mean, it depends on what on what you're looking for at one level. It's beautiful. Just if it works at all, another level, it's beautiful. If it's, if it uh, can be understood easily. It's beautiful. If it, uh, if it's a literate programming is beautiful. It makes you laugh. I mean, </s><s>[INST] yeah, I'm actually so I'm with you. I think Beauty. If it's readable, [/INST] read a book </s><s>[INST] is if you understand what's going on and also understand the elegance of thought behind it. And then also, as you said, wit and humor, I was always uh, remember having this conversation, I had this conversation on stack overflow. Whether humor is good in comments and I think it is [/INST] whether it is good in comments. </s><s>[INST] Like when you add comments code, I I always thought a little bit of humor is good. It shows personality. It shows character shows wit and fun and all those kinds of things of the personality of the program. [/INST] Okay, so, uh, a couple days ago I received a wonderful present from my former editor at Aston Wesley, he's downsizing his house and he found uh, that somebody had, the company had had found all the all of their internal files about the art of computer programming from the 19 sixties. And they gave it to him. Uh and then, uh, you know, before throwing throwing in the garbage and then so he said, oh yeah, he planned to keep it for posterity, but now he realized that posterity is a bit too much for him to handle. So he sent it to me. Uh and so, and so I just received this big, big stack of, of letters, uh, some of which I had written to them, but many of which they had written to early guinea pigs who were telling them whether they should publish or not, you know, And and one of the things I was uh in the in the comments to volume one, uh uh, the major the major reader was, was bob Floyd, uh who is my great uh coworker in the sixties, um, died early unfortunately. But uh and and he he commented about the humor in it. So we had, you know, he ran it by me, you know, says, keep this joke in or not, you know, uh you know, they also send it out to focus group, what do you think about humor in a in a book about computer program. And I stated my philosophy is it said, you know, the ideal thing is uh that it's something where the reader knows that there's probably a joke or if you only understood it and this is a motivation to understand to think about it a little bit. But anyway, it's a very delicate humor. I mean, it's really uh each each century invents a different kind of humor to, I mean, different different cultures have different, different kinds of humor. Um </s><s>[INST] Yeah, like we talked about Russia a little bit off line. You know, there's dark humor, you know, when a country goes to something different that [/INST] live and stuff like this and you and Jack Benny, I mean, steve Allen wrote this book about humor and it was the most boring book, but he was one of my idols, but but uh it's called the Funny men or something like that. But yeah, okay. So anyway, I think it's important to know that that this is part of life and and and it should be fun and not. And so, you know, I wrote this this organ composition which uh based on the bible, but I didn't refrain from putting little jokes and also in the music </s><s>[INST] getting in the music. It's it's [/INST] it's there. Yeah, </s><s>[INST] a little humor is okay. [/INST] Yeah. I mean, not egregious. You are so in this correspondence, you know, there were there were things I said, yeah, I really shouldn't have, I really shouldn't have done that. But other ones I yo insisted on. And I've got Jokes in there that that nobody has figured out. In fact in volume two, I've got a cryptogram, a message in Seaford. And in order to decipher it, you're going to have to have to break in our s a key which is larger than people know how to break. Uh So, you know, if computers keep getting faster and faster than, you know, might be 100 years, but somebody will figure out what this message is and they will laugh. I mean, I've got a joke in there. Uh </s><s>[INST] So that one you really have to work for. Uh I don't know if you've heard about this, Let me explain it. Maybe you'll find it interesting. So, open Ai is a company that does a. I work and they have this language model. It's a neural network that can generate language pretty well. But they also on top of that developed something called open ai codex. And together with Git hub, they developed a system called Open Ai co pilot. Let me explain what it does. There's echoes of literate programming in it. So what you do is you start writing code and it completes the code for you. So for example, you start let's go to your factoring program, you start you right in javascript and python in any language that are trained on uh you start you write the first line and some comments like what this code does and it generates the function for you and it doesn't incredibly good job, like it's not probably right, but often does a really good job of completing the code for you [/INST] to see whether, but how do you know whether it did a good job or not? </s><s>[INST] You can see a lot of examples where it did a good job, and so it's not a thing that generates, because for you, it starts, it gives you uh so it puts the human in the seat of fixing [/INST] issues </s><s>[INST] versus writing from scratch. Do you find that kind of idea at all interesting? [/INST] Every year we're going to be losing more and more control over what machines are doing and people are saying, well, it seems like when I was a professor at caltech, uh in the in the 60s, we had this this guy who who talked a good game, he could give inspiring lectures and you think, well, uh throwing things, he was talking about an hour later, you said, well, what did he say? Um uh, but he really felt that it didn't matter whether computers got the right answer or not, it's just a matter whether it made you happy or not. In other words, if you know, if your boss paid for, uh, you know, then you had a job, you could you could you could take care of </s><s>[INST] your wife, happiness is more important than truth. [/INST] Exactly. He didn't believe in truth, but he was a philosopher. </s><s>[INST] I like it. Uh, and somehow, you see, uh, [/INST] we're going that way. I mean, so many more things are taken over by saying, well, this seems to work. And when there's when there is a competing interests involved, neither side understands why the decision is being made. Uh you know, we realize now that it's that it's bad, but but consider what happens 55 year, 10 year, you're down the line when things get even more further detached. Each thing is based on something from the previous year. </s><s>[INST] So you start to lose, the more you automate, the more you start to lose track of some deep human essentially exponentially. But so that's the dark side. The positive side is, the more you automate the more you let humans do what humans do best. So maybe programming this, you know, maybe humans should focus on a small part of programming that requires that genius, the magic of the human mind and the mess you let the machine generate. Yeah, I mean, that's the that's the positive, but of course it does come with the darkness like automation. [/INST] What's </s><s>[INST] better, correct? [/INST] I'm never going to try to write a book about that. Uh I'm never going to recommend to any of my students to work for them. You're </s><s>[INST] on the side. I'm on the side. [/INST] I'm on the side of under understanding. I understand. Uh, and I think these things are really marvelous if they if if what they do is, you know, all of a sudden we have a better medical diagnosis or or help guide some scientific experiment or something like this. Uh you know, curing diseases or what whatever. But when it, when it affects people's lives in a serious way. Uh So if you're writing, if you're writing code, oh yeah, here this is great, uh this will make a slaughter. But okay. So </s><s>[INST] I see. So you have to be very careful. Like right now it seems like fun and games. It's useful to write a little javascript program that helps you with the website. But like you said, one year passes, two years passes five years and you forget you start building on top of it and then all of a sudden you have autonomous weapons systems based on [/INST] what we're all dead. Doesn't matter. And that's it. Well, </s><s>[INST] in the end, this whole thing ends anyway. So um but there [/INST] is a heat death of the universe a predicted But I'm trying to postpone that for for a little bit, </s><s>[INST] we'll be nice that at the end as we approached the heat death of the universe, there's still some kind of consciousness there to to to appreciate it. Hopefully human consciousness, [/INST] I'll settle for 10 to the 10 to the 10 to the 10th year. There's some finite number. But things like this might be the reason we don't pick up any signals from extraterrestrial, </s><s>[INST] they don't want anything to do with us. Oh because they because [/INST] they invented it to and </s><s>[INST] so you you do have a little bit of worry on the existential threats of ai and automation. So like like removing the human from the picture. [/INST] Yeah. People have more more potential to to do harm now than by far than they did 100 years ago. </s><s>[INST] But are you optimistic about the humans are good at creating destructive things, but also humans are good at solving problems. [/INST] Yeah, I mean, there's half empty and half full, you know, So </s><s>[INST] so how would we have full? I [/INST] can go. Yeah. So let me let me put it this way because because it's the only way I can be optimistic, but think of uh things that have changed because of civilization, you know, they don't occur just in nature. So just uh just imagine the room we're in, for example. Uh Okay. Some, you know, we've got pencils, we've got books, we've got tables, we've got microphones, clothing, food, all these things were added. Somebody invented them one by one, millions of things uh that we inherit. Okay. Um and uh it's inconceivable that that so many millions of billions of things wouldn't have problems and we get it all right. Um and and each one would have no negative uh effects and so on. So it's very amazing that much works as does </s><s>[INST] work. It, it's incredibly amazing. And actually that's the source of my optimism as well, including for artificial intelligence. So we we drive over bridges, we uh we use all kinds of technology. We don't know how it works and there's millions of brilliant people involved in building a small part of that and it doesn't go wrong and it works and I mean that it works, it doesn't go wrong often enough to suffer [/INST] and we can identify things that aren't working and try to improve on them </s><s>[INST] in a sub often sub optimal way. Absolutely. But [/INST] the the kind of things that I know how to improve require human beings to be rational and I'm losing my confidence that human beings are rational. </s><s>[INST] Yeah. Yeah. Now here you go again, with the worst case uh worst case analysis. Uh they may not be rational, but there um they're they're clever and uh beautiful in their own kind of way. I tend to think that most people um have the desire and the capacity to be good to each other and love will ultimately went out like if if they're given the opportunity, that's where they lean in the art of computer programming. You wrote the real problem is that programmers have spent far too much time worrying about efficiency in the wrong places and at the wrong times, premature optimization is the root of all evil in parentheses, or at least most of it in programming. Can you explain this idea? What's the wrong time? What is the wrong place for optimization? [/INST] So first of all the word optimization, I started out writing software uh and optimization was I was a compiler writer. Seo optimization meant uh making the uh making a better translation so that it would run faster on a on a machine. So an optimized program is just like, you know, you you want a programme and you set the optimization level. Uh the compiler, so that's one word for optimization. Um And at that time I happen to be looking at an unabridged dictionary uh for some reason or other. And I came toward optimizes, what's the meaning of the word optimized and it says to view with optimism And you you look in Webster's dictionary English language in 1960s or the 1960s, that's what optimized me meant. Okay. Um now, so people started doing cost optimization are the kinds of things uh you know, whole sub fields of of uh algorithms are not economics and whatever are based on what they call optimization now. But to me, optimization, when I was saying that was saying uh changing a program to make it more attuned to the machine. And I found out that uh when a person writes a program, uh he or she tends to think that the parts that were hardest to write are going to be hardest for the computer to execute. So maybe have 10 pages of color but I had to work a week writing this page. I mentally think that when they computer gets to that page is going to slow down. Uh, oh, I don't understand what I'm doing better be more anyway. This is of course silly. But it's something that we, that we, that we don't know when we've got a piece of code, we don't know what, whether the computer is actually going to be executing that code. Very much so. People have had a very poor understanding of, of what the computer was actually doing. I made one test where we studied a FORTRAN compiler And it was spending more than 80% of its time reading the comments card. Um, but as a programmer, we were really concerned about how fast it could take a complicated expression that had lots of blood levels of parenthesis and convert that into something. But that was just, you know, less than 1% of the uh, so if we optimize that, Uh, we didn't know what we were doing, but if we knew that it was spending 80% of his time on the comments card, you know, in 10 minutes, we could, we could make the compiler run more than twice as fast </s><s>[INST] and you can only do that once you've completed the program. And then you empirically study where [/INST] I had some kind of profiling that I knew what was important. </s><s>[INST] So you don't think this applies generally. I mean, there's something that rings true to this across. I'm glad [/INST] that it applied generally, but it was it was only my good luck I said it but you know but but I did but I said it in a limited context or not and I'm glad if it makes people think about stuff because I'm but it applies in another sense to that is um sometimes I will do optimization in a way that does help the the actual running time but makes the program impossible to change next week because I've changed my data structure is something that made it less adaptable. So one of the great uh principles of computer sciences is it his laziness or whatever you call it? Uh late binding uh you know, don't hold hold off decisions when you can um and you know, and we understand now uh quantitatively how valuable that is. </s><s>[INST] What do you mean? We understand. So you mean [/INST] People people have written thesis about how you can, how late binding or will improve the I mean, you know, just in time manufacturing or whatever you can make, you can defer a decision instead of doing an advanced planning and say I'm going to allocate 30% to this in 50 </s><s>[INST] percent. So in all kinds of domains there's an optimal itty to laziness in many cases, [/INST] decision is not made in advance. So instead you design in order to be flexible uh to change with the uh with the way the wind is blowing. </s><s>[INST] Yeah. But so the reason that line resonated with a lot of people is because uh there's something about the programmers mind that wants, that enjoys optimization. So it's a constant struggle. Two balanced laziness and lay binding with the desire to optimize the the elegance of a well optimized code is something that's compelling to programming. Yeah, [/INST] it's uh another concept of beauty. </s><s>[INST] Let me ask you a weird question. So, Roger Penrose has talked about computation computers and he proposed that the way the human mind discovers mathematical ideas is something more than a computer that that a universal turing machine cannot uh do everything that a human mind can do. Now, this includes discovering mathematical ideas and it also includes he's written a book about it, Consciousness. So I don't know if you know Roger, but do you think [/INST] my uh my daughter's kids played with his kids? And </s><s>[INST] so do you think there is such a limit to the computer? Do you think consciousness is more than a computation? Do you think the human mind the way it thinks is more than a computation? [/INST] I mean, I can say yes or no, but I don't think I have no reason. I mean, </s><s>[INST] so you don't find it useful to have an intuition in one way or the other. Like when you think about algorithms, isn't it? I think it's ananda limits [/INST] unanswerable question in my opinion is is no better than anybody </s><s>[INST] else. You think it's unanswerable. So you don't think eventually signed angels [/INST] can dance on the head of, I mean, I don't know, </s><s>[INST] but angels [/INST] uh anyway, there are lots of things that are beyond that we can speculate about, but I don't want somebody to say, oh yeah, can you said this? And and so he's smart. So that must be, I mean, I say it's something that we'll never know. Uh </s><s>[INST] interesting. Okay, that's a strong statement I don't I personally think is something we will know eventually. Like, there's no reason to me why the workings of the human mind are not within the reach of science. [/INST] That's absolutely possible. And I'm not denying it. Uh </s><s>[INST] but right now you don't have a good, I [/INST] mean, that's also possible, you know, that a I you know, created the universe, uh, intelligent design is all has all been done by an ai uh this is, I mean, all of these things are but but but but you're asking me to pronounce on it and I don't have any expertise. I'm a teacher that passes on knowledge, but I don't but I don't know the fact that I that I vote yes or no on. </s><s>[INST] Well, you do have expertise as a human, not as a not as a teacher or a scholar of computer science. I mean, that's ultimately the realm of where the discussion of human thought. Yeah, well, I know we're in consciousness. [/INST] I know where Penrose is coming from. He I'm sure he has no uh he might even thought he proved it, but </s><s>[INST] you know, he doesn't he doesn't prove he is following intuition, [/INST] but I mean you have to ask john McCarthy, john McCarthy uh I think uh we're totally unimpressed by these statements. Uh </s><s>[INST] so you don't think so. Even like the touring paper on on the touring tests that you know, starts by asking kIM machines think uh you don't think these kind of uh he touring doesn't like that question. [/INST] I don't consider it important, let's let's put it that way because it's in the category of things that it would be nice to know, but I think it's beyond knowledge. And so I don't, I'm more interested in knowing about the remind hypothesis or something. </s><s>[INST] So when you say it's an interesting statement beyond knowledge, I think what you mean is it's not sufficiently, well, it's not even known well enough to be able to formalize it in order to ask a clear question, so that's why it's beyond knowledge, but that doesn't mean it's not eventually going to be formalized. [/INST] Maybe consciousness will be understood someday. Uh the last time I checked, Uh it was still 200 years away, I haven't been specializing in this by any means, but but I went to lectures about it 20 years ago when I was uh there was, there was a symposium at the american Academy in in Cambridge and it started out by saying essentially everything that's been written about consciousness is is hogwash. Er </s><s>[INST] I tend to I tend to disagree with that a little bit so well, so consciousness for the longest time still is in the realm of philosophy. So it's just conversations without any basis and understanding still, I think once you start creating artificial intelligence systems that interact with humans and they have personality have identity, you start flirting with the question of consciousness not from a philosophical perspective, but from an engineering perspective and then starts becoming much more like, I feel [/INST] like you don't misunderstand me, I I certainly don't disagree with that at all. Um and I even at these lectures that we had, you know, 20 years ago there were neurologists pointing out that that human beings had actually decided to do something before they were conscious of making that decision. Uh I mean they could tell that, you know, that signals were being sent to their arms before they before they knew that they were uh things like this are true and and uh my less valiant has architecture for the brain and more recently, uh christmas Papademetriou in the Academy Science Proceedings a year ago with with two other people, but I know Crystal very well uh and and he's got this uh this model of uh this architecture by which you could create a things that that correlate well with uh with experiments that are done unconsciousness. Uh and any, he actually has a a machine language that in which you can you can write code and test hypothesis. Uh, and so it it might, you know, we might have a big breakthrough, my personal feeling is that consciousness the best model I I've heard of, uh, to explain the miracle of consciousness is that, uh, that somehow inside of our brains, we're having a uh, continual survival for the fittest competition. I'm speaking to you all the possible things I might be wanting to say. Are </s><s>[INST] you going? [/INST] Yeah. Right. Yeah. And one of them is winning. And and that's affecting, you know, the next sentence and so long, uh, there was this book, Machine Intelligence </s><s>[INST] on Intelligence [/INST] on Intelligence. Yeah. Bill Atkinson was was was a total devotee of that </s><s>[INST] book. I like whether it's consciousness or something else. I like the storytelling part that we it feels like, uh, first humans, it feels like there's a concrete, it's almost like literally programming. I don't know what the programming going on on the inside, but I'm getting a nice story here about what happened and it feels like I'm in control and I'm getting a nice clear story, but it's also possible there's a computation going on that's really messy. There's a bunch of different competing ideas. And in the end, it's just kind of generates a story for you to uh, a consistent story for you to believe. And that makes it all nice. [/INST] Yeah. And so I prefer to talk about things that I have some expertise, and then then things which which I'm only a uh on the sideline. </s><s>[INST] So there's a tricky thing, I don't know if you have any expertise in this, you might be a little bit on the sideline would be interesting to ask though, what are your thoughts on cellular automata in the Game of Life? Have you ever played with those kind of little games? [/INST] I think the Game of Life is his wonderful. And I uh and and shows all kinds of stuff about how things can evolve without the creator understanding anything more than the power of learning in a way. But to me, the most of important thing about the Game of Life is, is how it focus for me, what what it meant to have free will or not, because the game of Life is obviously totally deterministic. I find it hard to believe that anybody who's ever had Children cannot believe in free will. Right. On the other hand, this makes it crystal clear. Uh john Conway said uh he wondered whether it was immoral to shut the computer off after he got into a particularly interesting play of the Game of life. Um, wow, </s><s>[INST] Yeah. So there is to me the reason I love the Game of Life, it is exactly as you said, a clear illustration that from simple initial conditions with simple rules, you know, exactly how the system is operating is deterministic. And yet if you let yourself uh, if you allow yourself to lose that knowledge a little bit enough to see the bigger organisms that emerge and then all of a sudden they seem conscious. They seem not conscious, but living if [/INST] the universe is finite, we're all living in the game of life to slow down. I mean, it's sped up a lot. Uh </s><s>[INST] but do you think technically some of the ideas that you used for analysis of algorithms can be used to analyse the game of life? Can we make sense of it? Or is it too weird? And [/INST] yeah, I mean, I've got I've got a dozen exercises in volume for Fast ical six that actually worked rather well for that purpose. Bill gossipers came up with the algorithm that allows allowed Golly Too uh you know, to run thousands and thousands of times faster. You know the website called Golly T. O. L. L. Y. </s><s>[INST] Just, it stimulates the sell your item in a game of life. [/INST] Yeah. You got you got to check it out. </s><s>[INST] Can I ask you about john Conway [/INST] yes. In fact, I'm just reading now the the issue of mathematical intelligence or that came in last last week. And the whole issue devoted to uh you know, remembrance of him. </s><s>[INST] Did you know him? [/INST] I slept overnight in his house several times. I yeah, </s><s>[INST] he recently passed away. [/INST] You know, he he died a year ago. May I think it was of covid, </s><s>[INST] what are you, what are some memories of him of his work that stand out for? You did uh on a technical level that any of his work inspire you? On a personal level that did he himself inspire you in some way [/INST] and you know, absolutely to all of those things. But let's see, when did I first meet him? I guess I first met him at Oxford in 1967 when I was </s><s>[INST] okay. That's a long time ago. [/INST] Yeah. You were -20 years old or something, I don't know, 1967. But but uh, there was a conference where uh, I think I've spoken, I was speaking about something that no one has the canoes. Bendix algorithm now, but but he he gave famous talk about knots and and at the, and I didn't know at the time, but anyway, that talk now uh the stories of thousands and thousands of papers since since then. Uh and it was he was reported on something that he had done in high school, uh you know, almost 10 years earlier, uh, before this conference, but he never published it. And anti climax to stalk by building some not. So you have these these little plastic things that you that you can stick together. Uh It's something like lego but easier. Uh, so he made a whole bunch of, not in front of the audience and so on. And then disassembled. It was dramatic lecture before. He had learned how to give even more dramatic lectures later. So, All right. And </s><s>[INST] were you at that [/INST] lecture? And I was there, Yeah, because I had, you know, I was at the same conference um for some reason I was I happened to be in in Calgary uh at the same day that he was visiting Calgary And it was the spring of 72, I'm pretty sure. And and we had lunch together and he wrote down during the lunch on a napkin, all of the facts about what he called numbers. Uh and I and uh and he covered the napkin with the theorems about his his idea of numbers. And I thought it was incredibly beautiful. Um and and later in 1972, my sabbatical year began and I went to Norway and in december of that year, uh in the middle of the night, the thought came to me, you know, Conway's theory about numbers would be a great think teach students how to invent research and what the joys are of research. And and and so I said, and I had also read a book and dialogue, Bye bye Alfred renew. Uh We're kind of a Socratic thing where they're two characters were talking to each other about mathematics. And so I and so at the end in the morning I woke up my wife and said, Jill, I think I want to write a book about Conway's theory. And um you know, you know, I'm supposed to be right in the heart of computer programming, doing all this other stuff. But I got, but I really want to write this other book. And so we made this plan, but I said I thought I could write it in a week and we made the plan then. So in january, I rented a room in a hotel in downtown Oslo were on sabbatical in Norway uh and I ran at the hotel in downtown Oslo and did nothing else except right up Conway's theory. And I changed the name to surreal numbers. That so this book is now published a surreal number. And, and you know, we figured out, we always wondered what would you like to have a fair enough hotel rooms. So, so we figured out that she would visit me twice during the week. Things like this. You know, we would, you know, try to sneak in. This was hotel was run by a mission organization. These ladies were probably very strict, but anyway, so, so uh in the wild </s><s>[INST] week in every way. [/INST] But the thing is I had lost that. I had lost that napkin in which he wrote the theory, but I looked for it, but I couldn't, so I tried to recreate from memory what he told me. He had that lunch uh in Calgary. And and as I, as I wrote the book, I was going through exactly what I what the characters in the book we're supposed to be doing. So I start with the with the two axioms that start out the whole thing and everything is defined flows from that, but you have to discover why and and as every mistake that I make as I'm trying to discover it, uh my characters make to, you know and and and and so it was just a long, long story, but I work through this week uh and it was it was one of the most exciting uh intense weeks of my life and uh I described it in other places. But anyway, uh after six days I I finished it and on the seventh day, arrested and and I sent uh my secretary to type it. It was flowing as I was writing it uh faster than I could think almost. Uh but after I finished uh and try to write a letter to my secretary telling her how to type it, I couldn't write anymore uh to give it ever the muse had left me completely. Uh can </s><s>[INST] you explain how that we could have happened? Like why that seems like such a magical week. I have [/INST] no idea. But anyway, there was something, it was almost as if I was channeling. So so the book, it was typed, they sent it to Conway and and he said, well done, you got the accident, one accident wrong. Uh There is a difference between um less than or equal and not greater than I don't know the opposite of being greater than and less than or equal. But anyway technically it can make a difference when you're developing a logical theory and the way I had chosen was harder to to do than john's original. So um and we visited him at his house in Cambridge in april, we took a boat actually from Norway over to across the channel. And and so I didn't stay with him for some days. And and uh he told, he talked, we talked about all kinds of of things. He has, he had puzzles that I'd never heard of before. He had a great way to to solve the game of solitaire, many of the common interest that we, you know, he had never written up. But anyway, uh then in the summertime I took another week off and went to a place in in the mountains of Norway and rewrote the book using the correct accent. And so that was the most intensive connection with with with Conway after that. </s><s>[INST] Uh started with a napkin. [/INST] It started with a napkin, but but but we would run into each other. Uh Yeah, the next really and I was giving lectures in Montreal. Uh giving a series of Of of seven lectures about the topic called stable marriages. And and and he arrived in Montreal uh between my 6th and 7th lecture and and we met at a party and I started telling him about the topic I was doing and uh he sat and thought about it. He came up with a beautiful theory to show that the uh I mean in technical terms it's it's that the that the set of all stable marriages, it forms a lattice. And and there was a simple way to find the greatest floor bound and of two stable pairings and least upper bound to stable marriage. And so I could use it in my lecture the next day and he came up with this theme, you know, during the party. Uh it's a distributed last, I mean it's uh you know uh it added greatly to the theory of of stable magic. </s><s>[INST] So you mentioned your wife Jill much a stable marriage, Can you tell the story of how you two met? [/INST] So we celebrated 60 years of wedded bliss uh last month. And and we met because uh I was dating her roommate, this this was my sophomore year, her freshman year, I was dating her roommate and I wanted her advice on strategy or something like this. And anyway I found I enjoyed her advice better than enjoyed a roommate. </s><s>[INST] You guys were made during the same thing? No, no, no, because because I read something about working on a computer in grad school on a difficult computer science topic. [/INST] So so she's an artist and I'm okay and I'm uh geek and </s><s>[INST] what was she doing with a computer science book. Alright. I read, was that the manual that she was reading? [/INST] What was she wrote the manual that she had had, she had to take a class in computer science. Okay. And you're </s><s>[INST] the tutor. No, [/INST] no. Yeah. No we yeah we there were terrible times. Uh you know trying to learn certain concept, but I learned our premier and so we worked together uh you know occasionally in design projects, but every year we write a christmas card and and we each have to compromise our own notions of beauty. Uh </s><s>[INST] When did you fall in love with her [/INST] that day that I asked her about her roommate? </s><s>[INST] I mean, [/INST] no, okay. So you I don't mind telling these things depending on how you how far you go. But let me but let me tell you this that I I never really enjoyed kissing uh until I found how she did it, wow! </s><s>[INST] In 60 years. Is there a secret you can uh you can say in terms of stable marriages of how you stayed together so long. [/INST] The topic, stable marriage, by the way is not is a technical term. Uh </s><s>[INST] It's a joke done. [/INST] But uh to different people will have to learn how to compromise and work together and and you're gonna have ups and downs and crises and so on. And so as long as you don't Set your expectation on having 24 hours of bliss. Uh then there's a lot of hope for stability. But if you if you decide that it's that, that there's gonna be no frustration. Uh, </s><s>[INST] so you're gonna have to compromise on your notions of beauty when you write christmas cards, That's it. Uh, you, uh, you mentioned that Richard Feynman with someone you looked up to? Yeah, uh, probably you've met him in caltech. [/INST] Well, we knew each other. Yeah. At caltech for sure. Yeah. </s><s>[INST] Uh, you are one of the seminal personalities of computer science. He's one for physics. Have you, is there specific things you picked up from him by way of inspiration or? [/INST] So we used to go to each other's lectures and, but, but if I saw him sitting in the front row, I would throw me for a loop. Actually, I I would, I would miss a few, a few sentences. What unique story do I have about? I mean, I often referred to his, his time in brazil, uh, where he uh, essentially said they were teaching all the physics students the wrong way. They were just, they were just learning how to pass exams and not learning any physics. And he said, you know, if you want to prove it, you know, here, I'll turn to any page of this textbook and, and I'll tell you what's wrong with this page and he did so, and, and the text book had been written by his host. And, and, and it was a big embarrassing incident, but he had previously isis host if if he was supposed to tell the truth. Um But but anyway, it epitomizes the way uh education goes wrong uh in all kinds of fields uh and has to periodically be brought back from from from a process of giving credentials to a process of giving knowledge. </s><s>[INST] That's probably a story that continues to this day in a bunch of places where it's too easy for ah educational institutions to fall into credential is um versus uh inspirational is um [/INST] I </s><s>[INST] don't know if those are words but sort of uh understanding versus just giving a little um [/INST] it's </s><s>[INST] black [/INST] and you know, it's it's very much like what we're talking about if you want the computer too, If you want to be able to believe the answer, computer is doing one of the things Bob Floyd showed me in the 60s, there was a uh he loved this cartoon, there was there were two guys standing in front of, in those days, a computer was a big thing, you know? And and the first guy says to the other guy, he said, this machine can do in one second, what it would take a million people to do in 100 years? And the other guy says, oh, so how do you know? It's right, </s><s>[INST] it's a good line. Uh Is there some interesting distinction between physics and math to you? Have you looked at physics much to like speaking of Richard Feynman. So the difference between the physics community, the physics way of thinking, the physics intuition versus the computer science, the theoretical computer science, the mathematical sciences. Do you see that as a gap? Are they strongly overlapping? [/INST] It's quite different in my opinion. Uh I started as a physics major and I switched in the math uh and probably the reason was that I could I could get a plus on the physics exam, but I I never had any idea why I would have been able to come up with the problems that were on those exams. But but in math, I knew, you know, why the teachers set those problems and I thought of other problems that I could set to. And I believe it's it's quite a different mentality. </s><s>[INST] It has to do with your philosophy of geekdom [/INST] gigs. I mean some of my computer scientists friends are really good at physics and others are not. And I'm uh you know, I'm really good at algebra, but not a geometry, talk about different parts of mathematics, you know? So there are different kinds of physical, but physics just think of things in terms of waves and I can think of things in terms of waves, but it's like a dog walking on high legs if I'm thinking about. </s><s>[INST] So you basically, you like to see the world in uh in discreet ways and then more continuous. [/INST] Yeah, I'm not sure if during been a great physicist, I think it was a pretty good chemist. I don't know. But anyway, I see things I believe that computer science is largely driven by uh people who have brains, who who are good at resonating with certain kind of of of concepts and like quantum computers, it takes a different kind of </s><s>[INST] brain. Yeah, that's interesting. Yeah, it's well quantum computers is almost like at the intersection in terms of brain between computer science and physics because they involve both at least at this time. Uh but there is like the physicists have known they have incredibly powerful intuition [/INST] and there, I mean statistical mechanics. So I studied statistical mechanics and you know, I mean random processes uh are related to algorithms and a lot of a lot of ways, but there's lots of different flavors of flavors of physics as there are different layers of mathematics as well. Um but the thing is that I don't see well actually when they talked to physicists use a completely different language and when they're talking to, when they're writing expository papers, so I didn't understand quantum mechanics at all from reading about it in scientific american. But when I read, you know how they described it to each other, talking about Eigen values and and various mathematical terms, that that made sense. Then it made sense to me, but Hawking said that every formula you put in a book, you lose half of your readers. And so he didn't put any formulas in the book, so I couldn't understand his book at all, you could say you understood it, but I really I really didn't. </s><s>[INST] Uh Well the fireman also spoke in this way. So feynman, I think prided himself on a really strong intuition, but at the same time he was hiding all the really good the deep computation he was doing. So [/INST] there was one thing that that uh I was never able to uh I wish I had more time to work out with him, but I guess I could describe it for you there, There's something that got my name attached to it called Lucero notation, but it's a notation for very large numbers. And so uh I find out that that somebody invented it in in 1830s. Uh it's fairly easy to uh to understand anyway, so you start with X plus X plus X plus X and times and you can call that XN so X and his multiplication, then you take X times X times X times X. And end time. That gives you exponentially ation X to the power. So that's one arrow, X. So X N with no arrows, this multiplication X arrow and is X to the power </s><s>[INST] just to clarify for the uh so X times X times X and times is obviously X and X [/INST] plus X plus X and time. Uh </s><s>[INST] Oh yeah, okay. And then uh xn multiplication is X to the end. Uh and then and then here the arrow is when you're doing the same kind of repetitive operation for the exponential. [/INST] So I put in one arrow and I get extra power. Now I put in two arrows and that makes it takes X. To the X. To the X. To the X. To the ex N. Times priority. So in other words, if it's to uh double arrow three, that would be that would be to to the to to the two. So that would be two to the fourth power, that be 16. Okay, Okay. So so so that's the double arrow, and now you can do triple arrow. Uh of course. Uh and and so on. And and I had this this paper called uh well, essentially big numbers, uh you know, you try to impress your friends, but by saying a number they never thought of before. And and and I gave a special name for it uh Designed to font for it that has script K. And so on. But it really is 10, I think like 10 Quadruple Arrow three or anything like that. And I claim that that number if it is. So mind boggling that you can't comprehend how large it is. But anyway, fine, I talked to feynman about this and he said, oh let's just let's just use double arrow, but instead of taking editors, let's consider complex numbers. So, you know, you have, I mean okay, X X. Arrow arrow to that means ex ex uh ex but what about X X double arrow to 2.5. Well, that's not too hard to figure out that's interplay between the but what X double arrow, I or one plus I or some complex number. Uh And and uh so he claimed that there was no analytic function that would that would that would do the job, but I I didn't know how he could claim that. That was that wasn't true. Absolutely. And his next question was then have a complex number of arrows. Yeah. Okay. Okay. Okay. So that's that's fine. And uh violence. Yeah. </s><s>[INST] Can you describe what the uh ruth Morris pratt algorithm does and how did you come to develop it? One of the many things that you're known for and has your name attached to it? [/INST] Yeah. All right. So, it should be actually Morris pratt can youth, but we decided to use alphabetical order when we published the paper. The problem is uh something that everybody knows. Now, if there if they're using a search engine, uh uh you have a a large collection of text and you want to know if if the word can do with appears anywhere in the text to say or or some some other word that's less interesting than kind of. Okay. But </s><s>[INST] anyway, that's the most interesting or something. Right. [/INST] So we have we have a large piece of text and it's all one long, one dimensional thing, you know the 1st and 2nd letter etcetera, etcetera, etcetera. And so uh the question you'd like to be able to do this quickly. Um And the obvious way is uh let's say we're looking for Morris with them. Okay so we would we would go through and wait till we get to a letter m then we look at the next word and sure enough it said no and then they are but then what too bad um uh The next letter is E. So we missed we missed out on Morris and so uh we go back and start looking for another. Okay so that's the obvious way to do it. All right. Um And jim Morris noticed there was a more clever way to do it. The obvious way would have started let's say We found that that are in that character position 1000 it was started next step character position 1001. But but he but he said no look we we already read the O. And R. And we know that they aren't ems so we can we can start uh we don't have to read those over again. Uh So uh and this gets pretty tricky when when the word isn't Morris but it's more like abracadabra where you have patterns that are occurring uh </s><s>[INST] like repeating patterns in at the beginning at the middle, right, Right? [/INST] So uh he worked it out and he put it into the system software at Berkeley I think it was where he was uh he was writing some Berkeley UNIX, I think it was some routine, I was supposed to find occurrences of patterns in texas and yeah and yeah, we didn't explain it. And and and so he found out that several months later somebody had had looked at it didn't look right, and so they ripped it out. So he had this this algorithm, but it didn't make it through uh because what wasn't understood, nobody knew about this, particularly. Vaughan. Pratt also had independently discovered a year or two later. Uh I forget why I think Vaughn was studying some technical problem about palindromes or something like that. He wasn't really wan wasn't working on on text searching, but he was working on on an abstract problem that that was related. Well at that time, steve Cook was professor at Berkley. Uh and uh uh it was the greatest mistake that brookly CS department made was not to give him tenure. So steve went to went to Toronto, but uh but but I knew steve while he was at Berkeley and he had come up with a with a very peculiar theory uh by the technical concept called a stack of tom martin and a stack of tom martin, it is a machine that can't do everything a turing machine can do, but it can only look at something at the top of the stack or it can put more things on the stack or it can take things off the stack, like it can't remember a long string of symbols, but it can remember them in reverse order. So if you tell a stack of thomas on a B C D E. It can it can tell you afterwards. E D C B A. Uh you know, it doesn't have any other memory except except this one thing that it can see. And steve Cook proved this amazing thing that says, if a stack of time of time can recognize a language where the strings of the language or length end in any amount of time whatsoever to the stack of Tommy time, might use a zillion steps. Regular computer can recognize that same language in time and log in. So steve had a way of transforming a computation that goes on and on and on and on into using different data structures into something that you can do on a regular computer uh fast. The stack of Thompson goes slow, but somehow the fact that it can do it at all means that there has to be a fast way. So I thought this was a pretty, you know, cool theory. Uh And so I tried it out on on the problem where I knew a stack of Ramadan could do it, but I couldn't figure out a fast way to do it on a regular computer. I thought it was a pretty good programmer, but by golly, I couldn't think of any way to recognize this language uh efficiently. So I went through steve cooks construction, I filled my blackboard with all the everything that stack thomason dad did, you know, I wrote down and and then I tried to see patterns in that, and and how did he convert that into a computer program on a regular machine? Um and finally I psyched it out. What was, what was the thing I was missing so that I could say, oh yeah, this is what I should do in my program. Uh and now I have an official program. And and so I uh I would never have thought about that if I hadn't had his theorem, which was purely abstract thing. Well then I </s><s>[INST] used to try to into it how to use the stack of tom martin for the string matching problem. [/INST] Yeah. So, so the problem I I had started with with was not the string matching problem, but then I realized that the string matching problem was another thing which would also be, could be done by a stack of tomato. And and so when when I looked at what that told me, then I had a nice algorithm for this string matching problem, uh and it told me exactly what I should remember, as as I'm going through the string and I worked it out and and I wrote this little paper called automata theory can be useful and and the reason was that it was the first, I mean, I had been reading all kinds of papers about automata theory. Uh but it never taught me, it never improved my programming for everyday problems. Uh It was something that you published in journals and and you know, it was it was interesting stuff, but but here was a case where I couldn't figure out how to write the program. I had a theory from atomic theory, then I knew how to write the program. So this was uh for me uh a change in life, I started to say maybe I should learn more about thomas. And and I showed this note to bond prayer and he said he that's similar to something I was working on. Um And then uh and Jim Morris was at at Berkeley to at the time anyway, he's had an illustrious career, but I haven't kept track of Jim, but one is my colleague at stanford and and my student um later, but this was before Von one was still a graduate student and hadn't come to stand for you. So we found out that we had all been working on the same thing. So it was our algorithm we each discovered independently, but each of us had discovered a different, a different part of the elephant, uh you know, different aspect of it. And so we could put our things together. It was my job to write the paper, </s><s>[INST] how did the elephants bring to life [/INST] spring to life. Was because I I had drafted this paper, Atomic theory can be useful which was seen by Von and then by jim and then then then we combined because maybe they had also I've been thinking of writing something up about it, </s><s>[INST] about specific [/INST] between my problem in a period. </s><s>[INST] Let me ask a ridiculous question. Uh last time we talked, you told me what the most beautiful algorithm is actually. Uh for strongly connected graphs. What is the hardest problem puzzle idea in computer science for you personally that you had to work through. Just something that was just a [/INST] the hardest thing that I've ever been involved with? Okay, well yeah, that's I don't know how to answer questions like that, but in this case uh it's pretty clear because it's uh called the birth of the giant component. Okay, so now let me explain that because this actually gets gets into physics to and it gets into something called Bose Einstein Statistics. But anyway, it's got some interesting stories and it connected with Berkeley again. Um So start with the idea of a random graph. Now this is, did you hear we just say we have endpoints that are totally unconnected and and there's no geometry involved, there's no saying some points are further apart than others. All points are exactly are exactly alike. And Let's say we have 100 points and uh and we number them from 00-9 nine. All right, now let's let's take pi uh the digits of pi So two at a time. So we had 31 41 59 26. We can go through pie. And so uh so we take the 1st 2 31 41 let's let's put a connection between 410.31 point 41. Uh That's an edge in the graph, so Then we take 5926, make another edge. Uh the graph gets bigger, gets more and more connected as we add these things one at a time. Okay. We started out with endpoints and we add uh m edges. Okay, each edges completely, we forgot about it and we had before we might get edge twice, we might get edge from a point to itself. But um you know, maybe Pie is going to have a run of four digits in there, so we're gonna but anyway we're evolving a graph at random. Um And uh a magical thing happens when the number of edges is like 0.49 and Uh huh maybe it is a million. And I have uh you know, 490,000 edges. Uh then almost all the time it consists of isolated trees, not even any loops. Uh </s><s>[INST] It's a very small number of veggies so far, [/INST] A little less than half in. And but if I had .51 edges, so a little more than half in so it's a million points. 510,000 edges. Now, it probably has uh a one component that's much bigger than the others. Mhm. Um And we call that the giant component. </s><s>[INST] Can you clear? So can you clarify? So I thought is there a name for this kind of random supercool pie random graph? [/INST] Well I call it pot pie graph. No no I I the pie graph is actually my pie graph is based on binary representation of pine. Not the decimal representation of pi but anyway let's suppose I was rolling dice instead. So </s><s>[INST] I might have so it doesn't doesn't have to be [/INST] part of any source of the point. Is every step choose totally at random. One of those endpoints choose totally at random another one of the endpoints make that an edge. That's the process. Yeah. </s><s>[INST] So there's there's nothing magical about pie. They you [/INST] know I was using pi to sort of saying pious sort of random that nobody knows a pattern in. Exactly. Uh but it's not yeah I could have just as well drawn straws or something. Um This was a concept invented by edition rainy and they called evolution of random graphs. And if you start out with with a large number and and you repeat this process all of a sudden the big bang happens at one half end there will be two points together then maybe we'll have three. Uh you know the then they maybe branch out a little bit but but I'll be separate until we get to one half end and we passed one half end and all of a sudden there's substance to it that there are, there's a big clump of stuff that's all joined together. </s><s>[INST] So it's almost like a phase transition of some [/INST] kind. It's exactly it it's a phase transition, but it's a double phase transition. It turns out it happens. I mean there's actually two things going on at once at this phase transition, which is uh which is very remarkable about. Okay, so, so uh a lot of the most important algorithms are based on random processes. And so I wanted to, you know, I want to understand random processes now. So there are data structures that sort of grow this way. Okay, so, so, so Dick Karp, one of the leading experts on random randomized algorithms, had his students working looking at this at Berkeley and we heard a rumor that the students have found something interesting happening. You know, the students are generating this, are simulated this random evolution of graphs and undertaking snapshots. Uh It was so often take a look at what the graph is. And the rumor was that every time they looked there was only one component that had loops in it. Almost always, they do a million experience. And uh only three or four times did they ever happened to see a loop at at this point? Uh no more than one component with the loop. So they're watching until the graph gets completely full. Uh So it starts out totally empty and gets more and more more and more edges all the time. Uh And and so Ok, certainly a loop comes along once, but now all the loops stay somehow joined to that one. There never there never were two guys with loops </s><s>[INST] uh [/INST] in these experiments. Okay, so anyway, this almost always. Not always, but but with very high probability that seemed to be true. So so we heard about this rumor at stanford and we said if that's true then must you know, a lot more must also be true. So there's a whole bunch, there's a whole theory out there waiting to be discovered that we have ever thought about. So let's take a look at that. And so we looked closer and we found out no, actually it's not true, but but in fact it's almost true. Namely there's a very short interval of time what is true and if you don't happen to look at during that short interval of time then you miss it. So in other words there will be a period where there two or three components have loops but they joined together pretty soon. </s><s>[INST] So [/INST] if you don't have a real fast shutter speed, you're gonna miss you're going to miss that instant. </s><s>[INST] So separate loops don't exist for long. [/INST] That's it. Yeah, I started looking at this to make it quantitative and uh the basic problem was to slow down the big banks so that I could watch it happening. I think I can explain it actually in fairly elementary terms. Even without writing a formula, like talking would do uh and and so uh that's a let's watch the evolution. And at first uh these edges are coming along and they're just making things without loops, which we call trees. Okay. So then all of a sudden the loop first appears. So at that point I have one component that has a loop. All right now, now I say that the complexity of a component is the number of edges minus the number of vertex is so if I have a loop, I have like a loop of length five has five edges and five vortices. Um or I could put a tail on that and that would be another edge. Another vertex, </s><s>[INST] 0, 1, 2 complexity kind of thing. So [/INST] if the if the complexity is zero, we have 11 loop, I call it a cycle or I call the cyclic components. So cyclic component looks like a wheel to which you attach fibers or trees, they go branching. But there's no more loops. There's only one loop and everything else feeds into that loop. Okay. And that has complexity zero. But the tree itself has complexity minus one because it has uh like like like it might have 10 versus season and nine edges to tie the time together. So nine minus 10 is minus one. So complexity -1 is a tree. It's got to be connected. That's what I mean by a compound. That's got to be connected. So, so, if I have 10 things connected, I have to have nine edges. </s><s>[INST] Can you clarify why when complexity goes? Uh can go about zero? I'm a little. Yes. [/INST] What? Right, so the complexity plus one is the number of loops. So, if complexity is zero, I have one loop. If if complexity is one, that means I have one more edge than I have heard it. So, I might have like 11 edges and 10 vortices. So it turns we call that a bicycle because it's got two loops and it's gonna have two loops. And if well, </s><s>[INST] why can't you be trees just going out for the loop? [/INST] That I would need more edges than All </s><s>[INST] right, okay, So [/INST] every time I get another loop I get another excess of edges over verte asi Ok, so in other words, uh we start out and after I have one loop, I have one component to ask a cycle in it. Now, the next step, uh according to the rumor would be at the next step, I would have a bicycle. Uh in the evolution of almost all graphs, it would go from cycle two bicycle, but in fact, there's a certain probability it goes from cycle to to, you know, 2 2 different cycles. All right. Um and I worked out the probability was something like five out of 24, there was it was substantial. Uh but still uh soon they're going to merge together almost. Okay. So but then it splits again after you have either Either two or 1 one. Uh the next step is you either have three or you have to one or you have 111. Okay. And so I worked out the probability for those transitions and I worked it out up to up to the first five transitions. And I had these had these strange numbers 5 24 and I stayed up all night and about three a.m. I had the numbers computed and I looked at him and here we're the denominator was something like 22, So the probability was something over two, </s><s>[INST] I don't know how you work that out. But I had a [/INST] formula that I could calculate the probability and I could find the limiting probability as n goes to infinity. And it turned out to be this number. But the denominator was too thrilled. And I looked at the denominator, I said, wait a minute. This number of factors because 1000 and one is equal to seven times 11 times 13. I had learned that my first computer program. So, so, so, so 23 0 23 is seven times 11 times 13 times 23. Uh That's not a random number. There has to be a reason why those small crimes appear in the denominator, but my thing. So all of a sudden that suggested um, another way of looking at the problem where small prime factors would occur. </s><s>[INST] So what would that be? So [/INST] that said, oh yeah, let me take the law algorithm of this formula. And sure enough, it's going to simplify and it happened so and I wouldn't have noticed it except for this factory ization. Okay. So I go to bed and I say, ok, this looks like I'm flowing down the big bang. I can figure out what's going on here. And the next day it turned out Bill Gates comes to stanford to visit. Uh they're trying to sell him on donating money for a new computer science building? And, and, and uh, they gave me an appointment to talk to Bill and I and I wrote down on the blackboard this, this evolutionary diagram, you're going from 1 to 25 24th and all this business. And I wrote it down and anyway, at the end of the day, the uh, he he was discussing people with the development office and he said, boy, I was really impressed with what professor can you said about this giant component. And uh, and so, you know, I, I love this story because it shows that theoretic computer science is really worthwhile. You know, </s><s>[INST] does Bill, have you ever talked to Bill Gates about it? Uh since then. Yeah, that's cool. That's cool in a moment in [/INST] history. But anyway, he happened to visit. I'm exactly the day after I had I found this pattern and and that allowed me to crack the problems so that I could develop the uh theories are more and understand what's happening in the big but uh because I could I could now write down explicit formulas for stuff. So it would you work not only the first few steps but also though study the whole process and and I worked further and further and I with two authors co authors. And we finally figured out That the probability that the rumor is true in other words, look at the evolution of uh of a random graph going from 00 to to complete and say what's the probability that at every point in time there was only one component with the cycle. Mhm. We started with this rumors saying there's only one like there's only one component with the cycle and </s><s>[INST] uh 100 [/INST] Percent Rumor was that 100%. And it turned out the actual numbers like 87%. I should remember the number but I don't but I don't have a with me. But anyway, But but but the number, it turned out to be like 12 over pi squared or anyway, so eight over anyway. It it was a nice it related to pi um and we could never have done that with. So that's the hardest problem I ever saw in my life was to prove that this probability is </s><s>[INST] it was proven. The probability was proven. [/INST] Yeah, I was able to prove this, that this and this should shed light on a whole bunch of other things about random graphs. That was sort of the, the major thing we were after after </s><s>[INST] that. Super. Cool. I what was the connection of physics that you mentioned? Well, [/INST] Bose Einstein statistics is the study of how molecules uh bond together uh without geometry. Without this x </s><s>[INST] you created the tech typesetting system and released it is open source just on that little aspect. Why did you release it as open source? What is your vision for open source? [/INST] Okay, well that the word open source didn't exist at that time, but but I didn't want proprietary rights over Because I saw how proprietary rights were holding things back. In the late 50s, people at IBM developed the language called FORTRAN. They could have kept it proprietary, they could have said on the IBM can use this language. Everybody else has to but they didn't, they said, anybody who can write, who can translate for training into the, into the language of their machines is allowed to make FORTRAN compilers to um on the other hand, in the topography industry, I had seen a lot of languages that were developed for composing pages and each manufacturer had his own language for composing pages and that was holding everything back because people were tied to a particular manufacturer. And and then a new equipment has invented a year later. But printing printing machines, they have to Expect to advertise the cost over 2030 years. </s><s>[INST] So you didn't want that for tech. [/INST] I didn't need the income. I already I already had a good job. And you know, uh my books were uh people were buying enough books that I that that it would bring me plenty of supplemental income for everything my kids needed for education, whatever. So there was no reason for me to try to maximize income. Any further income is sort of a threshold function. If you don't have, if you don't have enough, you're starving. But if if you get over the threshold, then you start thinking about philanthropy or else you're trying to take it with you. But uh but anyway, there's a I my income was over the threshold, so I didn't need to keep it. And so I specifically could see the advantage of uh making it open for everybody. </s><s>[INST] Do you think most software should be open? [/INST] So I think that uh people should charge for non trivial software, but not for trivial software. </s><s>[INST] You give an example of, I think Adobe Photoshop versus Gimp on Lennox as Photoshop has value, [/INST] which so it's definitely worth paying paying for all this stuff. I mean, I mean, well they keep adding adding stuff that's that my wife and I don't care about but somebody, but I mean, but they have built in a fantastic uh a new feature, for example, in Photoshop where you you can go through a sequence of 1000 complicated steps on graphics and it can take you back anywhere in that sequence uh with really beautiful algorithm. I mean, yeah, it's </s><s>[INST] Oh, that's interesting. I didn't think about what algorithm it must be some kind of efficient representation [/INST] really? Yeah, no, I mean there's a lot of really subtle Nobel prize class creation of intellectual property in in there. Um and uh and with patents you've got a limited time to uh I mean, eventually the idea of patent is that you publish so that it's not secret is not a trade secret </s><s>[INST] that said you, you said that I currently use a bond to Lenox on a standalone laptop. It has no internet connection. I occasionally carry flash memory drives between the machine and the max that I use for network surfing and graphics. But I trust my family jewels only to Lennox. Why do you love Lennox? [/INST] The version of politics that I use is stable. I actually, I'm going to have to upgrade one of these days but </s><s>[INST] to a newer version of a bond to [/INST] Yeah, I'll stick with you want to, but right now I'm running something that doesn't support a lot of the new Software. It's the last stable re I don't remember the number, we're like 14 anyway. It's quite and I'm going to get a new computer, I'm getting new solid state memory instead of instead of a </s><s>[INST] hard disk. Well, let me ask you uh sticking on the topic of tech um when thinking about beautiful typography, what is your favorite letter number or symbol? Mhm. I know, I know, ridiculous question. But is there some or? Uh huh. [/INST] Look at the last page. Mhm. The very end of the index. Mhm. </s><s>[INST] Uh huh. What is that? [/INST] There's a book by Dr Seuss called and Beyond Zebra and he gave a name to that. </s><s>[INST] Did you see Dr Seuss give a name to that. [/INST] Dr Seuss, this is S. P. U. S. S. He he wrote Children's books In the 50s, 40s and 50s. </s><s>[INST] You're talking about cat in the hat. [/INST] Cat in the hat. Yeah, that's it. Yeah, I like how you have on on Beyond Seabra. Did you get to the soviet union? </s><s>[INST] Yeah, Dr know Dr Seuss did not come to the soviet Union, but since you Oh actually, I think it did actually a little bit when we're um that that was a book his maybe cat in the hat or or green eggs and ham I think was used to learn english. So I think it made it in [/INST] that way. My okay, I didn't like those as much as Bartholomew covens, but I used to know. Bartholomew covers my heart when I was young. </s><s>[INST] So what the heck is this symbol? we're looking at. There's so much going on. [/INST] He has a name for it at the end of his book on Beyond Zebra </s><s>[INST] who made it? He did, he did. So there's it looks like a bunch of vines. Well, is that simple existence, [/INST] By the way, he made a movie in the early 50s. uh I don't remember the name of the movie now, you can probably find it easily enough. It features uh dozens and dozens of piano is all playing together at the same time. And but but all the scenery is sort of based on the kind of artwork that was in his books and uh the fantasy, you know, based of Sousse Land or uh and I saw the movie only once or twice, but it's but it's quite, I'd like to see it again. </s><s>[INST] That's that's really fascinating that you gave them. They gave him a shout out here. Okay, is there some elegant basic symbol that you're attracted to, some uh give something that gives you pleasure? Something used a lot [/INST] pi pi of course, uh try to use pi uh as often as I can when I need a random example, because it doesn't have any uh known characters. So for for instance, I don't have a care to show you, but uh do you know the uh the game called Mass? Um A S. Y you it's it's a great recreation. I mean, Sudoku is easier to understand what Moscow is. it is more addictive. Uh You have black and white stones like a like on a go board. Uh And you have to draw a path that goes straight through a white stone and makes a right angle turn out the Blackstone. Um And it turns out to be really really nice puzzle because It doesn't involve numbers, what visual, but it's three D. Pleasant to play with. So I wanted to use it as example in art of computer programming. And I have um I have exercise on how to design cool Masu puzzles, you can find out on Wikipedia certainly uh as an example M. A. S. Y you. Um And and so I and so I decided I would take pi the actual image of it and you had pixels and I would put a stone wherever it belongs in the letter pie in the greek letter pi. And but the problem was find a way to make some of the stones white some of the stones black so that there's a unique solution to the Matsu puzzle. That was a good test case for my algorithm on how to design Moscow puzzles because I insisted in advance that the stones had to be placed in exactly the positions that make a letter pi make a letter. All right. And and I saw, you know, and it turned out there was a unique way to do that. Um And so so pie is a source of of examples where I can, where I can prove that I'm starting with something that isn't can uh And most most recently I was writing about something called graceful graphs. Yeah uh graceful grass is the following. You have a graph that has uh m edges to it. And you attach numbers to every vertex in the following way. So every time you have an edge between vortices, you take the difference between those numbers and and that difference is he's got to be tell you what edge it is. So that one edge, two numbers will be one apart. There'll be another edge where the numbers are two apart. And so uh great computer problem. Can you find a graceful way to label a graph? Uh So I started with, so I started with a graph that I use for organic graph? Not not a Mathematically symmetric graph or anything. And I take this, take a 49 states of the United States. Uh The edges that go from one state to the next state. So for example, California be next to Oregon, Nevada Arizona. Okay. And and and I include district, District of Columbia. Uh So I have 49 I can't get Alaska and hawaii in there because they don't touch, you have to be able to drive from one to the other. So is there a graceful labelling of of the United States, Each state gets the number and then if California is number 30 and Oregon is number 11, that edge is going to be number 19. The difference between those. Okay, so is there a way to do this for for other States? And uh and so I was I was thinking of having a contest uh for people to get it as graceful as they could. Uh But my friend tom rookie actually solved the problem by proving that, I mean, I was able I was able to get it down Within seven or something like that. He was able to get a perfect solution </s><s>[INST] the actual solution or to prove that a solution exists [/INST] more precisely. I had figured out a way to put labels on so that All the all the all the edges were labeled somewhere between one and 217, But but there were some some some gaps in there because I should really have gone from 1 to 105 or what whatever the number is. So I gave myself too much, a lot of slack. He did it without any slack whatsoever. Perfect, graceful labelling. And and so I call out the contest because the problem is already solved and too easy instance, because tom was able to do it in the afternoon. Um </s><s>[INST] he's sorry, he gave the algorithm or for this particular uh for [/INST] the United States, the United States, this problem is this problem is incredibly hard. I mean for the General </s><s>[INST] general is uh but [/INST] It was very lucky that we worked for the United States I think, but the theory is still very incomplete. But anyway, then tom came back a couple of days later and he had been able to not only find a graceful labelling but but the label of Washington was 31. The label of Of Idaho was 41. Following the digits of Pi. Going across the topic of the United States. He has the digits </s><s>[INST] of pi perfectly on purpose. [/INST] He was able to still get a graceful labelling with that with that extra thing. Uh It's a miracle. Okay, but I like to use pie in my book, you see? And this is the uh </s><s>[INST] all roads lead to pi somehow uh somehow often hidden in the middle of like the most difficult problems. Can I ask you about productivity? [/INST] Productivity? </s><s>[INST] Yeah. You said that, quote my scheduling principle is to do the thing I hate most uh on my to do list by weeks and I'm very happy. Can you explain this process to a productive life? [/INST] Oh I see. Well, but all the time I'm working out on what I want, what I don't want to do, but still I'm glad to have all those unpleasant task finished. </s><s>[INST] Yes. Is that something you would advise to others? [/INST] Well, I yeah, I don't know how to sit well during the pandemic, I feel my productivity actually went down by half because I have to um I have to communicate by writing which is slow, I have to I mean I don't like to send out a bad sentence, so I, you know, I go through and reread what I've written and read it and fix it. So everything takes a long longer when I'm when I'm communicating by my text messages, uh instead of just, you know, together with somebody in the room. And it's also slower because the libraries are closed and stuff. But there's another thing about scheduling that I learned from my mother that I should probably tell you and that is um it's different from what people in robotics field do, which is called planning. So she had this principle that was see something that needs to be done and do it. I would just, instead of saying I'm gonna do this person do this first, just, you know, uh just do it, pick this up, but </s><s>[INST] you're at any one moment there's a set of tasks that you can do and you're saying a good heuristic is to do the one you want to do least [/INST] right. The one I haven't got any good reasons, I will never be able to do it any better than I am now. Uh I mean there are some things that I know if I do something else first, I'll be able to do that one better, but there's some that are going to be harder because you know, uh I've forgotten some of the ground work that went into it, or something like that. So I just finished a pretty tough part of the book, and uh and so now I'm, you know, doing the parts that are more fun, but but but the other thing is, as I'm writing the book, of course, I went the reader to think that I'm happy all the time. I'm writing the book. I'm, you know, that it's upbeat. I I can have humor, I can, you know, I can I can say this is cool, you know? Well, in this, uh I have to I have to disguise the fact that it was painful in any way. The </s><s>[INST] road to that excitement is painful. Yeah, it's laden with pain. Okay, is there you've given some advice to people before? But can you, uh [/INST] can you give me too much credit? But anyway, this is my this is my turn to just uh to say things that that I believe, But but I want to preface it by saying, I also believe that other people do. How did these things much better than I do? So, I can only tell you my side of it. </s><s>[INST] So, can I ask you to give advice to young people today to high school students, to college students, whether they're geeks or the other kind about how to live a life there. It can be proud of how to have a successful career, how to have a successful life. [/INST] That's it's always the same as I've said before. I guess that not to do something because because it's trendy, but it's something that you personally feel that you were called to do rather than somebody else expects you to do. </s><s>[INST] How do you know, you're called to do something? [/INST] You tried and it works or or or it doesn't work. I mean, you learn about yourself, life is a binary search, you try something and you find out, oh yeah, I have a background that helped me with this, or maybe I'm, maybe I could do this if I worked a little bit harder, but you try something else and you say I have really no intuition for this. And it looks like uh you know, it looks like it doesn't have my name on, it </s><s>[INST] was their advice along the way that you got about what you should and shouldn't work on, or do you just try to listen to yourself? [/INST] Yeah, I probably overreact another way when, when something, when I see everybody else going some way, I probably, I probably say that's too much competition. Uh, but, but but uh, but mostly I I played with things that were interesting to me and then later on I found, oh, actually the most important thing I learned was how to be interested in almost anything. I mean, not to be born her, it makes me very sad when I, when I see kids talking to each other and they say that was boring and to me, a person should feel upset if he would help if he had to admit that he wasn't able to find something interesting. </s><s>[INST] Uh, [/INST] you know, the skill they saying, I haven't learned how to, uh, how to enjoy life. I have to have somebody entertain me instead of. </s><s>[INST] That's really interesting. It is a skill. Uh, David Foster Wallace. I really like the thing he says about this, which is the key to life is to be unbearable. And I do really like you saying that it's a skill because I think that's a really good, that's really good advice, which is if you find something boring, that's not, I don't believe is because it's boring. It's because you haven't developed [/INST] how </s><s>[INST] to find the beauty and how to find the fun in it. That's a, that's a really good point. [/INST] Sometimes it's more difficult than others. I have to do this. But I mean, during the covid lots of days when I never saw another human being, uh, but I still find other ways to, </s><s>[INST] it still was a pretty fun [/INST] time. Yeah, I came earlier. I came a few minutes earlier today and I walked around foster, so I didn't want, you know, I didn't know what's going on in foster city. I saw a beautiful, some beautiful flowers at the nursery at home depot for a few blocks away. </s><s>[INST] Life is amazing. It's full of amazing things like this. Yeah. I just, sometimes I'll sit there and just stare at a tree, nature is beautiful. Uh Let me ask you the big ridiculous question. I don't think I asked you last time I have to ask this time in case you have a good answer. What is the meaning of life? Our existence here on earth? The whole thing. Yeah. Uh No, no, you can't. You can't I will not allow you to uh to try to escape answering this question. You have to answer definitively because [/INST] there </s><s>[INST] surely surely don canoe. There must be an answer. [/INST] What is the answer? Is it 42 or </s><s>[INST] I don't think it's a numerical. That's that [/INST] was that was in Shannon. Okay. But all right. So talk to anyway. It's only for me. And but I but I personally think of my belief that that God exists, although I have no idea what that means. But I believe that there is uh some something beyond human uh capabilities. Um, it might be uh might be some ai uh but but whatever it is, but what whatever I but I do believe that that there is uh something that goes beyond the realm of human understanding. But but that that I can try to learn more about how to resonate with whatever that being would like me to do. </s><s>[INST] So you think you can have occasional glimpses of that being. [/INST] I strive for that. Not that I ever think I'm going to get close to it. But it's not, it's not for me it's saying, what should I do? That that being wants me to do. That's that's, you know, I'm trying to ask what that means. Is that being want me to to be talking to lex Friedman right now, You know? And I said yes, okay, but thank you. Well, thank you. But what I'm trying to say is I'm not trying to say what of all this Strategies I could choose or something. Which 1? I try to do it. Not, not strategically, but I try to to imagine that I'm following somebody's wishes </s><s>[INST] even though you're not smart enough to know what they are. Yeah, but that funny little dance. [/INST] Well, I mean, there's A I or whatever it is probably is smart enough to help to give me clues. Uh, </s><s>[INST] And uh, you make the whole journey from clue clue a fun one. [/INST] Yeah. I mean, it's as so many people have said, it's the journey, not the destination and people live live through crises help each other. I think things come up. Uh, history repeats itself. Uh, you try to say in the world today. Is there any government that's working? I read history. I I know that things were, oh, they </s><s>[INST] were there were there were a lot worse in many ways. [/INST] There's a lot of bad things all the time. And I read about. Uh, you know, I look at things that people have good ideas and they were working on great projects. And then I know that it didn't succeed though in the end? But the new inside have gotten him actually in that way, was I was reading, what book was I reading recently? It was it was by ken Follett and it was called The Man from ST Petersburg. But but it was talking about the prequel to World War and Winston Churchill, according to this book, uh, seems that that Germany has been spending all its gold reserves, building up a huge military. And there's no question that if Germany would attack England, that England would be wiped out. Um, so he wants Russia to help, uh, to attack Germany from the other side, because Germany doesn't have enough of an army to be fighting two wars at one. Okay, now, then there's an anarchist in Russia who sees that wars are, uh, something that leaders start. But actually people get killed. And so he wants to stop any alliance between England and Russia, because that would mean that thousands and thousands of people of Russia would would would be killed. That wouldn't be otherwise killed. All right. And so his life's goal is to assassinate a Russian prince who is visiting England, because that will make you mean the Zero will not form the alliance? All right. So, we have this year question about what should the government do? Should it actually do something that will lead to, you know, is the war inevitable or is there a way to have people? And it struck me that if I were in a position of responsibility for people's lives, in most cases I wouldn't have, I wouldn't have any confidence that any of my decisions were good. That that that that these questions are too hard probably for any human being, but certainly for me. </s><s>[INST] Well, I think I think coupling the not being sure that the decisions are right. So that that's actually a really good thing, coupled with the fact that you do have to make a decision and carry the burden of that. And ultimately I have faith in human beings and the great leaders to arise uh and help build a better world. I mean that's the hope of democracy. That's the Ben. [/INST] Let's hope that we can enhance their abilities with uh with algorithms. </s><s>[INST] Uh well put done. It's such a huge honor. You've been an inspiration to me and to millions for such a long time. Um thank you for spending your really valuable time with me once again, it's a huge honor. I really enjoyed this conversation. Thanks for listening to this conversation with Donald Knuth to support this podcast. Please check out our sponsors in the description and now let me leave you with some words from Don Canute himself. Science is what we understand well enough to explain to a computer art. Is everything else we do. Thank you for listening. I hope to see you next time. [/INST] </s> |
<s>[INST] you thinking also when you say open source at the hardware level, is this going to the design architecture instruction set level? Or is it going to literally the, you know, the manufacturer of the of the actual hardware, of the actual chips, whether that's a six specialized in particular domain or the general. [/INST] Yeah. So let's talk about that a little bit. So when you get down to the bottom layer of uh software, the way software talks to hardware is in a vocabulary. And what we call that vocabulary we call that uh the words of that vocabulary called instructions. And the technical term for the vocabulary is instruction set. So those instructions are like what we talked about earlier, there can be instructions like add subtract and multiply divide. There's instructions to put data into memory Which is called the store instruction and to get data back, which is called the load instructions. And those simple instructions uh go back to the very dawn of computing and you know in 1950 the commercial commercial computer had these instructions. So that's the instructions that we're talking about. So up until I'd say 10 years ago, these instruction sets are all proprietary. So a very popular one is old and by Intel, the one that's in the cloud and then all the PCS in the world. Intel owns that instruction set. It's referred to as the X- 86. There have been a sequence of ones that The first number was called 80 86. And since then there's been a lot of numbers but they all end in 86. So there's been that uh kind of family of instruction sets </s><s>[INST] and that's proprietary. [/INST] That's proprietary. The other one that's very popular is from arm that kind of powers all the, all the cell phones in the world, all the ipads in the world. And a lot of things that are so called internet of things, devices arm. And that one is also proprietary Arm will license it to people for a fee, but they own that. So the new idea that got started at Berkeley uh kind of unintentionally 10 years ago is uh early in my career, re pioneered a way to do of these vocabulary instruction sets. That was very controversial at the time. at the time in the 1980s, conventional wisdom was these vocabularies, instruction set should have, you know, powerful instructions. So polysyllabic kind of words you can think of that. And and so that instead of just add subtract and multiply they would have polynomial vied or sort a list. And the hope was of those powerful vocabularies that make it easier for software. So we thought that didn't make sense for microprocessors. Service people at Berkeley and stanford and IBM who argued the opposite. And we called that was a reduced instruction set. Computer in the abbreviation was ri sc. And typical for computer people we use the abbreviation start pronouncing it. So risk. So we said for microprocessors which with Gordon's more is changing really fast. We think it's better to have a pretty simple a set of instructions, reduced set of instructions. That would be a better way to build microprocessors since they're going to be changing so fast due to moore's law. And then we'll just use standard um uh software to cover the you generate more of those simple instructions. And one of the pieces of software that is in the software stack going between these layers of abstractions is called a compiler and it's basically translates. It's a translator between levels. We said the translator will handle it. So the technical question was uh well since they're these reduced instructions, you have to execute more of them. Yeah, that's right. But maybe you can execute them faster. Yeah, that's right. They're simpler. So they could go faster, but you have to do more of them. So what's what's that trade off look like? And it ended up that we ended up Executing maybe 50% more instructions, maybe a third more instructions but they ran four times faster. So so this risk controversial risk ideas proved to be maybe factors of three or four better. Um I love that this idea was controversial </s><s>[INST] and almost kind of like rebellious. So that's in the context of what was more conventional is the complex instruction set computing. So how do you pronounce that? Sisk? [/INST] Sisk risk versus Sisk? And uh and believe it or not, this sounds very, very uh who cares about this? Right. It was it was violently debated at several conferences, is like what's the right way to go is is and people thought risk was, you know, was de evolution. We're going to make software worse by making Those instructions simpler. And their fierce debates at several conferences in the 1980s And then later in the 80s, that kind of settled uh to these benefits. </s><s>[INST] It's not completely intuitive to me why risk has for the most part [/INST] one? </s><s>[INST] So [/INST] why did that happen? </s><s>[INST] Yeah, yeah. And maybe I can sort of say a bunch of dumb things that could lay the land for [/INST] further commentary. So uh to me this </s><s>[INST] is uh this is kind of interesting thing if you look at people as possible to see with modern compilers, you really could write faster code with C plus plus. So relying on the compiler to reduce your complicated code into something simple and fast. So to me comparing risk. Maybe this is a dumb question. But why is it that focusing the definition, the design of the instruction set on very few simple instructions in the long run, provide faster execution. Versus coming up with, like you said, a ton of complicated instructions than over time, you know, years, maybe decades. You come up with compilers that can reduce those into simple instructions for [/INST] you. Yeah, So let's try and split that into two pieces. So if the compiler can do that for you. If the pilot can take a complicated program and produce simpler instructions, uh then the programmer doesn't care, right, programmer, I don't care just how how fast is the computer I'm using, How much does it cost. And so what we what happened kind of in the software industry Is right around before the 1980s, critical pieces of software were still written not in languages like C or C plus plus. They were written in what's called assembly language where there's this kind of humans writing exactly at the instructions at the level than that a computer can understand. So they were writing add subtract, multiply instructions. It's very tedious. But the belief was to write this lowest level of software that that people use, which are called operating systems. They had to be written in assembly language because these high level languages were just too inefficient, they were too slow or the programs would be too big. So that changed with a famous operating system called UNIX, which is kind of the grandfather of all the operating systems today. So UNIX demonstrated that you could write something as complicated as an operating system in a language like C. So once that was true them, that meant we could hide the instruction set from the programmer. And so that meant then it didn't really matter. The programmer didn't have to write lots of these simple instructions that was up to the compiler. So that was part of our arguments for risk, is if you were still writing assembly language, there's maybe a better case for cis constructions, but if the compiler can do that, it's going to be uh you know, that's done once the computer translates it once and then every time you run the program it runs at this, uh this potentially simpler instructions. And so that was the debate, right? Is because people would acknowledge that the simpler instructions could lead to a faster computer. You can think of monosyllabic constructions, you can say that, you know, if you think of reading, you probably read them faster or say them faster than long instructions, the same thing. That analogy works pretty well for hardware and as long as you didn't have to read a lot more of those instructions, you could win. So that's that's kind of that's the basic idea for risk. </s> |
<s>[INST] and, and even if you think you can ignore it, just reading it will have a little bit of an effect on your mind. I think that's, I think that's right over time, I might accumulate. So the documentary, but in general, uh, it mentioned or kind of emphasized and talked about erin's sexuality or sort of, they were discussing basically the idea that he was a homosexual and some of the trauma, some of the suffering that he endured his life had to do with sort of fear given the society of, of what his father would think of, what others around him, especially in sports culture and football and so on. So, I don't know, in your interaction with him, was do you think that maybe even leading up to a suicide, do you think his struggle with coming to terms of the sexuality had a role to play in much of his difficulties? [/INST] Well, I'm not going to talk about my interactions with them and anything I derived from, from that, but you know, what I will say is that, um, a story broke on the, on the radio, uh, at some point during the trial that Aaron had been in the same sex relationship with someone and some sport, local sportscasters, local boston sportscasters are we really mushroomed, uh, the story. So, um, uh, he and everyone, I was aware of it. Um, you'll, you also may know from the court record that the prosecutors floated a, uh, specious theory, um, for a minute, but then backed off of it, that, you know, that Aaron was that there was some sort of uh, I guess gay rage at work with them and that might be a cause a motive for the killing. And uh, luckily they really backed off of that. That was quite an offensive claim and in theory. So, um, but to answer your question more directly, I mean, I have no idea why he killed himself. It was a surprise and a shock. I was scheduled to go see him like a couple of days after it happened. I mean, he was anxious uh, for uh, Jose and I to come in and do the appeal from the murder which he was convicted for. He wanted us to take over that appeal. Um, He was talking about going back to football. I mean, he said, well you talked about this earlier. You talked about the sort of innocent aspect of him. He said, you know, well, Ron maybe not, maybe not the patriots, but you know, I want to get back in the league. And I was like, you know, Aaron, that's, that's gonna be tough man. Uh, but uh, but he really, you know, he really believed it. And uh, and then, you know, for a few days later that to happen. It was just, it was a real shock to me. </s><s>[INST] Like when you look back at that at his story, does it make you sad? [/INST] Very, very, I thought so, one I believe he absolutely did not commit the crimes that we could acquitted him on. I think that was the right answer. Uh for for for that. Uh I don't know enough about Bradley, the first case. I'm sorry to make a make an opinion on, but in our case, uh you know, uh it was just he had the misfortune of family, a famous name and the police department just really just just just really got got on him there. So um uh yes, it's it's it's I miss him a lot. It was very very sad, surprising. </s><s>[INST] Yeah. And and I mean just on the human side, of course, we don't know the full story, but just everything that led up to suicide, everything right up to uh an incredible professional football player, you know, that whole story if [/INST] remarkably talented athlete, remarkably talented athlete. </s><s>[INST] And it has to do with all the all the possible trajectories, right, that we can take through life as we were talking about before. And some of them lead to uh to suicide sadly enough. And it's always tragic when you [/INST] have some </s><s>[INST] you know, somebody with uh with great potential result in and the things that happened. People love it. When I ask about books, I don't know if uh uh whether technical like legal or fiction nonfiction books throughout your life have had an impact on you. If there's something you could recommend or something you could speak to about something that inspired ideas, insights about this world complicated world of ours. [/INST] Oh wow. Uh yeah, so uh I'll give you a couple. Uh so one is a contingency irony and solidarity by Richard Ward E. He's he's passed away now, but was a uh philosopher at some of our major institutions, Princeton, Harvard stanford, um Contingency irony in solidarity, at least. That's the book that really helped me work through uh a series of thoughts. So it stands for the proposition that uh that are most deeply held beliefs are contingent, that there there's nothing beyond history or prior to socialization. That's definite story of the human being that's already. Um and he says that our most deeply held beliefs are received wisdom and highly contingent along a number of registers. And he does that. But then goes on to say that uh he nonetheless uh can hold strongly held beliefs recognizing their contingency but still believes them to be true and accurate. And it helps you to work through uh what could be an intellectual tension uh other words. So, so you don't delve into one, doesn't delve into relativism, everything is okay, but it gives you a vocabulary to think about uh how to negotiate these these these realities. Uh </s> |
<s>[INST] and it's the right thing to do. Have you seen in history, [/INST] there are examples where it's possible to pause the Moloch? Absolutely. Human cloning, for example, you could make so much money on human cloning. Why aren't we doing it? Because biologists thought hard about this and felt like this is way too risky. They got together in the 70s and the Selimar and decided even to stop a lot more stuff, also just editing the human germline, gene editing that goes into our offspring, and decided, let's not do this because it's too unpredictable what it's gonna lead to. We could lose control over what happens to our species. So they paused. There was a ton of money to be made there. So it's very doable, but you just need a public awareness of what the risks are, and the broader community coming in and saying, hey, let's slow down. And another common pushback I get today is, we can't stop in the West because China, and in China, undoubtedly, they also get told, we can't slow down because the West, because both sides think they're the good guy. But look at human cloning. Did China forge ahead with human cloning? There's been exactly one human cloning that's actually been done that I know of. It was done by a Chinese guy. Do you know where he is now? In jail. And you know who put him there? Who? Chinese government. Not because Westerners said, China, look, this is... No, the Chinese government put him there because they also felt they like control, the Chinese government. If anything, maybe they are even more concerned about having control than Western governments, have no incentive of just losing control over where everything is going. And you can also see the Ernie Bot that was released by, I believe, Baidu recently. They got a lot of pushback from the government and had to rein it in, you know, in a big way. I think once this basic message comes out that this isn't an arms race, it's a suicide race, where everybody loses if anybody's AI goes out of control, it really changes the whole dynamic. It's not, I'll say this again, because this is a very basic point I think a lot of people get wrong. Because a lot of people dismiss the whole idea that AI can really get very superhuman, because they think there's something really magical about intelligence such that it can only exist in human minds. You know, because they believe that, I think it's gonna kind of get to just more or less GPT four plus plus, and then that's it. They don't see it as a suicide race. They think whoever gets that first, they're gonna control the world, they're gonna win. That's not how it's gonna be. And we can talk again about the scientific arguments from why it's not gonna stop there. But the way it's gonna be is if anybody completely loses control and, you know, you don't care if someone manages to take over the world who really doesn't share your goals, you probably don't really even care very much about what nationality they have. You're not gonna like it. It's much worse than today. If you live in Orwellian dystopia, what do you care who's created it, right? And if someone, if it goes farther, and we just lose control even to the machines so that it's not us versus them, it's us versus it. What do you care who created this unaligned entity which has goals different from humans ultimately and we get marginalized, we get made obsolete, we get replaced. That's what I mean when I say it's a suicide race. It's kind of like we're rushing towards this cliff, but the closer the cliff we get, the more scenic the views are and the more money there is there. So we keep going, but we have to also stop at some point, right, quit while we're ahead. And it's a suicide race which cannot be won, but the way to really benefit from it is to continue developing awesome AI a little bit slower. So we make it safe, make sure it does the things that humans want and create a condition where everybody wins. The technology has shown us that geopolitics and politics in general is not a zero sum game </s><s>[INST] at all, so there is some rate of development that will lead us as a human species to lose control of this thing and the hope you have is that there's some lower level of development which will not allow us to lose control. This is an interesting thought you have about losing control. So if you are somebody like Sandra Pritchey or Sam Altman at the head of a company like this, you're saying if they develop an AGI, they too will lose control of it. So no one person can maintain control, [/INST] no group of individuals can maintain control. If it's created very, very soon and it's a big black box that we don't understand like the large language models, yeah, then I'm very confident they're gonna lose control. But this isn't just me saying it. Sam Altman and Demis Asabes have both said, they themselves acknowledge that there's really great risks with this and they wanna slow down once they feel it gets scary. But it's clear that they're stuck in this, again, Moloch is forcing them to go a little faster than they're comfortable with because of pressure from just commercial pressures, right? To get a bit optimistic here, of course, this is a problem that can be ultimately solved. To win this wisdom race, it's clear that what, we hope that it was gonna happen, hasn't happened. The capability progress has gone faster than a lot of people thought, and the progress in the public sphere of policymaking and so on has gone slower than we thought. Even the technical AI safety has gone slower. A lot of the technical safety research was kind of banking on that large language models and other poorly understood systems couldn't get us all the way, that you had to build more of a kind of intelligence that you could understand. Maybe it could prove itself safe, things like this. And I'm quite confident that this can be done so we can reap all the benefits, but we cannot do it as quickly as this out-of-control express train we are on now is gonna get the AGI. </s> |
<s>[INST] track. What do you think is the role of family and wrestling can a man do it alone and if not, where is family most important? [/INST] You know, you can do it alone, but why would you want to, I think the chances of doing alone are much less than the chances of doing it together. Yeah, I know they say don't bring your profession home. Sometimes they say that, but I never, I never got away from my profession and you know, sometimes I, it's like my house right here. So when I'm moving home, I'm not going to have an office because I'm not going to coach anymore or I'm not going to be an assistant athletic director for a while that you got to do something that it gives you a little bit of a break. Not you necessarily, maybe the person you're living with. And so I don't know if you looked outside there, I got a cabin right in my backyard. You probably can't see it right </s><s>[INST] there. But what's in the cabin? [/INST] That's my house away from my house. It's only 30 ft from my house and it's my office and it's my workout room. It's my, I got a son of there. It's, it's a bed upstairs if I need it, if I ever get too close. And she says, hey, why don't you go sleep in the other house? But you know, it kicks me out of the bed, but it's never happened. But I do spend a lot of time out there and uh, it's uh, you know, you gotta have a little distance sometimes, you got to know you're gonna know your role. And so all of a sudden when you're a guy that's been gone your whole life from eight o'clock in the morning until close to 7 30 or eight o'clock at night. So 11, 12 hours a day. And then all of a sudden you're not gone as much even though you still work, she's trying to slow me down now. I'm doing, uh, not so much like here what we're doing right now, but when I get in the car and drive somewhere or fly somewhere, you know, like just last night I went to bed and I hadn't told her that this guy called me and he wants me to, uh, speak for, uh, you want to build another wrestling, wants to start another wrestlers and business networking out and uh, Delaware because we don't have any colleges and wrestling in Delaware. And so I said, well, you know, I'm gladly that because that's my life, you know, so, but then all of a sudden, I didn't say that to my wife until all of a sudden this morning and I told her that I might go on the friday the 21st of december. </s><s>[INST] Oh no, [/INST] well, well I said that's not christmas, she goes, we're celebrating christmas that weekend early because a lot of the family can't be here except for that weekend. And I said, oh, well, that's not gonna work, but I kind of didn't say anything to her first. And then, well, I'll tell you she started getting a little emotional And if I wanna stay married for another year, 45 years, then I better tell those people that I got family obligations because uh that depends what's most important. I love wrestling, I love wrestling and I want to start another, help start another rustlers and business network, but there's more than one day and gable out there. Well, maybe not, but there's there's a lot of people that are uh maybe even closer and they've got big names. I mean, we're doing pretty well right now. I mean, we got first two years ago and we got second this year, uh and then we got the women's freestyle is doing good in wrestling, we gotta work a little bit our Greco yet, but but they are working on it. But our men's freestyle team right now are are excellent and you know, and the key for them is to get them all on the same page instead of just have new highlights. And by that I'm saying who you look and see who won this year. Well, the three guys that have never won before one this year, we had three world champions are two past world champions. Didn't win this year. I mean they did okay. you know, they got medals. </s><s>[INST] The boroughs [/INST] win. No he did not. He got third. </s><s>[INST] Oh that's right, you got bronze. Yeah [/INST] july. I've got I mean Snyder got second. So those two are main guys, you know, so the three new guys that came through, we're guys that hadn't won world gold. In fact two of them have never made a world team before. And so when we have three world champions this year, but we needed all five of them to come through to win the championships. And so the key really is getting them all to do the same at the same time you're in and you're out and not just based on okay uh Burrows got beat to share so he'll win next year. It's gonna be every year if you're capable of doing that and that's what the coaching staff has to do. What's kind of funny that I do have a lot of influence actually on the coaching staff right now at the U. S. A. Level because the the women's freestyle guy is uh is terry Steiner and he wrestled for me was a national champion. He's got a twin brother that's at Fresno State. And then Billy's attic is the uh freestyle coach and he wrestled for the Hawkeyes back in uh the early days and he was the national champion. So we got a lot of former gable influence on there, but they've got deep roots in there </s> |
<s>[INST] So you don't know what the heck the meaning is. You just know you're a bit different than the others and then maybe, well maybe the whole point is you want to realize you want to let that madness flourish, that uniqueness flourish, that's the whole point of life. We're all different and are in like very interesting little ways. And the more different you are, you want to let that you want to let that become, you want to let it be as full. It's like a garden, you know, all the garden flower you did mention, um you weren't sure if there's a free will or not. Um Do you think it's all predetermined? Or do you think will be the choices? We make [/INST] our decisions? I just said if it is, but I know that we make our decisions and and I know that we are spirits that are living in this, in this flesh. That's beyond a shadow of a doubt with me. If you walk out of your body and have out of body experience and you will know it. </s><s>[INST] So the body is just the temporary, the spirit lives [/INST] on eternally with no beginning and no end. And that's hard to fathom. Yeah, this is just a little ship. This is a shell to contain that spirit, you know, This is the way we work on earth, You know? But yeah, I know I'm an eternal being. So are you </s><s>[INST] think there's a y to it, you know? Do you think there's a meaning to this life? [/INST] Well, I think the why is beyond my capability of understanding. It's someone greater than me. I don't understand it, but it's awesome. I just know that it's awesome. And one day we will know. The answer is once we get to that cross over the other to the other side, I think we will understand clearly. It says, you know now we see through a glass darkly. But then when we are face to face with God, we will understand. </s><s>[INST] And until we know, let's just enjoy this. Beautiful. Well we got a [/INST] man that was my guilt. I love everybody and everything I do. And it's just and I'm sorry if I put a stumbling block in anybody's way, I wouldn't want to. But uh these are these things and I just think about, oh, what a hypocritical world we live in though. Like uh most anybody listen. Okay, he's a drug dealer and I would say most of the committed adultery. That's a cardinal sin. Yet they moved, threw threw rocks at me for moving uh marijuana cocaine across the road. It's just if you saw the two different things you'd say what a terrible difference it is. But we become conditioned. But this mad society that we have </s><s>[INST] you mentioned that your daughter Miriam wrote you a poem. You mind reading it? [/INST] I'll be glad to I was doing 11 years up in lompoc penitentiary, maximum security prison for parole violation For possession of marijuana in 1977. They should have gave me six months. But they gave me 11 years because they wanted me for what they call silent beef Anyhow. While I was in that dungeon, I received a letter from my daughter Miriam Gold Daddy's poem. A year ago. I became a poet when I wrote your birthday pros. And here I am today. Ready to give it another go first. I would like to wish you a very happy birthday to be and to thank you so very much for without you, I would not be me. Secondly, I want to say that your support has been immense. It has been true, honest loving and free of all. Pretense. 3rdly, it goes without saying you love it surpassed all my wrongs and you always made me smile with one of your old country songs. I can remember on cuervo Daddy with you holding me in your arms as you sang Jim Reeves songs and talked about the farm. I can see you walking through the door from one of your travels far and wide. And the thought of you coming home. Daddy kept a twinkle in our eyes. I can smell you as I did when I used to climb into your bed and you would talk to me again about one of the adventures that you lead. I can see me and Mario asleep in one of your airplanes Extraordinaire. And remembering wondering myself why there wasn't an available chair. I remember having to meet you and worrying that you wouldn't be there. But you would pop up from behind some counter and give us all a happy scare. You gave us presence in key biscayne and hotels, pleasure galore and three dozen roses that we came through the airport door. I can see your face in Amsterdam with the luggage carousel and you look like a boy with a secret that you were just dying to tell. He taught me mathematics and the sands of far away places and taught me to seal. And we left without any traces. We climb glaciers in Argentina and saw the blue of the beautiful caves and witnessed the majestic beauty of such a jag ling mies. I learned how to change gears on the dirt roads of brazil. We eat hot dogs in Paraguay. A memory. We smile over still. We talked about lions, elephants, and bears on the hacienda and Uruguay, but decided it was better after europe. We did fly over the old world and all this luxury. What a good time! It was from south America to the cosmic Pawelski. I think we fell in love the european joint. Well, it is considered a book in itself, but it's a story about beauty and knowledge, suspense and worldly wealth. We went from holland to Sweden. We went from France to Spain and I promise you, I have no regrets. I would definitely do it all again. I would see the world with you anytime sir. There's no doubt in my mind because being by your side, Daddy always ensures a wild good time. So our past took a turn and we're back in the U. S. Of a life here isn't so bad. And I'm plum content to stay. I'm happy to be near you. Although I'm not as close as it was before because of your love and encouragement. I've been able to open your doors. I'm grateful to be in school and I'm generally happy where I am. And I even like when you call and tell me to study for the next exam, what a life you've given me, daddy. It's a tremendous and a magical gift. We already have so many stories to tell their far too many to list. But I want to thank you again this day with a very big happy birthday to you. And you tell you just a few more things that I knew in my heart to be true, that I love you daddy with all of your wrongs and your rights. That you're ahead of our family and you've kept us a whole bound tight that you have an honest love in your heart for God and all mankind and you truly do believe in yourself. When you say it will all be fine. I know you will be there to catch me if ever I waver a slip and I know I want jewish captain on any sinking ship. I also know a new chapter is written. It's almost time to move on. It's time to sail another C. and to witness a brand new dawn. It'll be good to see you at the helm again. As you point out our destination, they laugh and dance on the upper deck is while the boat glides through. It'll be good to see you on the go because I know you like to be and to know you can open any door without any key. But while we revel in our days together, we will know better than to hurry Because as you told me many times, life is an incredible journey. Wow, </s> |
<s>[INST] And so in your career, you walked amidst violence. But you never participated in the violence. [/INST] I didn't even see it. Yeah, it just didn't happen around me in prison. It did. I slowed people up. They called me doc. I dental floss and uh 11 time I had to get a blade and try to help keep them from my patient from getting again. But I don't, we're just like if I shot at those people I shot at him to keep him from killing me. I did I certainly didn't mean to kill them. So that's that's just some people are evil and they will kill you and hurt you lie to you. I just don't do any of that. It's just it makes you sick. I've seen it when I was in to shoot three guys tried to kill the guy and they stabbed him so many times. But they're stabbed blake and the blood getting out of the room and you're gonna kill him. You're going to kill him, save his life, drug him up there with the guards. Could see him. There's stuff like that. I'm just not of that nature of those people. They're just evil. There's people born evil. I believe </s><s>[INST] it is heartbreaking to hear that the basic humanity has gone in prison in the United States. That's heartbreaking because that basic humanity is actually the light at the end of the tunnel is the thing that saves us as opposed to uh when it's absent is the thing that destroys us. [/INST] The prisons are filled absolutely filled with people that have some mental problems. Now you see tent city all the way up and down here. I guarantee you every one of those people have mental problems some degree however little it is but they're a little bit off now then you get a D. E. A. Agent. It wants to make a name for himself. He goes down and gets two of them? One of them to sell a little two g of methamphetamine to the other one and he gets a conviction and a young prosecutor, he gets a conviction, he wants to make a judge and we got the judge and where is it? I'm gonna give a million what's his name? Gilbert. I'm gonna make I'm gonna give him in a million years before I get off the judge. You get fools like that in charge. You're gonna fill prisons up with beautiful humanity. And those are the ones and then the other is people over drugs in drugs should be a health issue. You can't you cannot police is enough. It's just they know like the only thing that overdoses is opioids heroin and if they can give it to him it costs about a dollar a day to give the worst addict is his fixed. But they'll give it methadone which is from a pharmaceutical company which is just as bad. Why in the world we tried it all over the world and uh. Portugal in England and when they give the girls clean up no more stolen cars, why who who wants to keep this forrest going? They just perpetuating it like oh every little police places. Getting all these suits and armor and machine guns. It's just like oh it's it's such a spin. It's it's sad. </s><s>[INST] Do you think all drugs should be legalized? [/INST] I don't know about that, but they certainly should be control. If a person is an addict, he should be able to go down and get his fix with somebody there to help him with a clean needle and a glass of orange juice. It's so much cheaper than prison. So much cheaper than him stealing cars or prostitute having to go to work. That's sad. </s><s>[INST] You've lived one heck of a life. Looking back there is, uh, there's a lot of young people that listen to this high school college students, what advice would you give them? How to live? How to have a successful career? I don't have a good life, how to be a good man, a woman [/INST] to be a good man or woman. If I had it to do over with, I'll just tell you what I'd have done. I would have paid attention and studied my lesson and did the best I could. Um, in school, in school, yes. And went as far as I could have, I would have liked to been a doctor. I just didn't have the stick ability or anybody to tell me, hey, go there and do that. And if you can do that at a very young age, starting a trade, learn to do something. It doesn't matter what it is. If you learn to do something good, there is a great demand for you. And I would say that in prison, the prison system should come in and you get a thief. Young fans of thief robber and you see all right, we need uh we need carpenters, We need plumbers, we need electricians. We need sheep sentence, sentence them to that trade. And when you get an A plus in that where you can go out and make you $30 or $50 an hour, you go home Now you can you mess around 10 years if you want to or you can do this in two. I think that would uh that's just for the prison. But anyway, I would say that they find somebody and be true to him that we have uh just be honest and true in your life. </s><s>[INST] You mean like relationships, relationships. [/INST] Yes. I mean, so many, so many people particularly our Children are from relationships where they're not wanted, their divorce, their fathers left, They don't know who the daddy is. They descend foster homes, 500,000 Children in foster homes in America today. And and we have our government inadvertently isn't encouraging those people. My daughter is a doctor and she delivered a couple years ago a baby From a 10 year old child. That child. And she said in the individual room is four generations, all of them on welfare. Now we got one more and it reminds me of Elvis Presley song in the ghetto. </s> |
<s>[INST] to push back, Is it possible that the crazier the person, the less likely they are and in a way where we're safe, meaning like this is the kind of proposition I've had, I had the discussion with my dad is a physicist about this, where he thinks that uh uh like if you have a graph like evil people can't also be geniuses. So his, this is his defense why evil people will not get control of nuclear weapons because to be truly evil, but evil meaning sort of, you can argue that not even the evil of Hitler we're talking about because Hitler had a kind of view of Germany and all those kinds of, there's like he probably deluded himself and the people around him to think that he's actually doing good for the world is similar with Stalin and so on. By evil. I mean more like almost like terrorists where they want to destroy themselves and of the world, like those people will never be able to be actually skilled enough to do to deliver that kind of mass scale destruction. So the hope is that It's very unlikely that the kind of evil that would lead to extinction of humans or mass destruction is so unlikely that we're able to last way longer than 700, years. [/INST] That's free. It's very unlikely in that sense. I accept the argument, but that's why you need to let the clock tick. It's also the best argument for bureaucracy To negotiate a bureaucracy, it actually select against pure evil because you need to build alliances. So bureaucracy in that regard is great, right? It keeps out the worst apples. But look, put it this way, could you imagine 35 years from now the Osama bin laden of the future has nukes are very bad bioweapons. It seems to me you can Yeah, and Osama was pretty evil and actually even he failed, right? But nonetheless, That's what the seven or 800 years is there for. </s><s>[INST] And it might be destructive technology that don't have such a high cost of production or such a high learning curve. Make cyber attacks or artificial intelligence, all those kinds of things. [/INST] Yeah, I mean, let me ask you a question, let's say you could as an act of will By spending $1 million dollars obliterate any city on earth and everyone in it dies and you'll get caught uh and you'll be sentenced to death. But you can make it happen just by willing it, how many months does it take before that happens. </s><s>[INST] So the obvious answer is like very soon. There's probably a good answer for that because you can consider how many millionaires there are. You can look at that, right? I have a sense that there's just People that have $1 million. I mean there's a certain amount, but have $1 million dollars have other interests that will outweigh the uh the interest of destroying entire city? Like there's a particular, you know, like maybe that's a hope that's why [/INST] we should be nice to the wealthy too, right? Yeah. </s><s>[INST] Yeah. All that trash talking is Bill Gates. We just stop that because uh that doesn't inspire the other future. Bill Gates is to be nice to the world. [/INST] That's </s><s>[INST] true. But your senses the cheaper guest to destroy the world, the more likely it becomes. [/INST] Now when I say destroy the world, there's a trick in there. I don't think literally every human will die, but it would set back civilization by an extraordinary degree. It's then just hard to predict what comes next. But a catastrophe where everyone dies that probably has to be something more like an asteroid or supernova. And those are purely exogenous for the time being at </s><s>[INST] least. So I immigrated to this country and I was born in the soviet union in Russia. And uh [/INST] which one? Which is an important question? Well, when you were born in the soviet union, </s><s>[INST] right, Yes. I was born in the soviet union. The rest of the details. But I grew up in Moscow Russia. But I came to this country and this country even back there, but it's always symbolized to me a place of opportunity where everybody could build like uh build the most incredible things, especially in the engineering side of things just invent and build and scale and have a huge impact on the world? And that's been to me the, that's my version of the american ideal of the american dream. Uh Do you think the american dream is still there? Uh Do you think, what do you think of that notion itself from an economics perspective? From a human perspective? Is it still alive? And how do you think about it? The american dream? [/INST] The american dream is mostly still there. If you look at which groups are the highest earners? It is individuals from India and individuals from Iran, which is a fairly new development. Great for them. Not necessarily easy. Both. You could call persons of color may have faced discrimination also on the grounds of religion. Uh yet they've done it. That's amazing. It says great things about America. Now, if you look at native born americans, the story is trickier. People think intergenerational intergenerational mobility has declined a lot recently, but it has not for native born americans For about I think 40 years, it's been fairly constant, which is sort of good. But compared to much earlier times it was much higher in the past. I'm not sure we can replicate that because look, go to the beginning of the 20th century, very few americans finish high school or even have much wealth. There's not much credential ism there aren't that many credentials. So there's more upward mobility across the generations </s><s>[INST] than today. [/INST] And it's a good thing that we had it. I'm not sure we should blame the modern world for not being able to reproduce that. But look, the general issue of who gets into Harvard or Cornell, is there an injustice? Should we fix that? Is there too little opportunity for the bottom say half of americans? Absolutely. It's a disgrace how this country has evolved in that way. And in that sense, the american dream is clearly ailing, but it has had problems from the beginning for blacks, for women from any other groups. </s> |
<s>[INST] Oh boy, that's dark. You're looking for major things in history that don't yet have a- Cognate going forward, right? [/INST] And that's a nice way to think about the future. It is only one model and, you know, any mental model or something like that. That's why I say them as a sci-fi scenario. It's just like a scenario one could contemplate, right? Where the new version has, I mean, the Chinese communists do not think of themselves as Aryans, right? But they are ultra-nationalist. And, you know, the Hitler comparisons, people talk about Hitler endlessly, you know, like Saddam is a new Hitler, everybody's a new Hitler, et cetera. If there is a comparison to, quote, Nazi Germany, it is, you know, CCP China in a sense, why? They are non-English speaking, manufacturing powerhouse, with a massive military build-out under one leader that is a genuine peer competitor to the US on many dimensions. And in fact, you know, exceeding on some dimensions of technology and science, right? That is like, the problem is it's a boy who cried wolf. People will say this a zillion times, right? And, you know, that is like, you know, I'm not saying this, by the way, crucially, I am just like, I think China is very complicated and there's hundreds of millions of people, probably half in China, that disagrees with the current ultra-nationalist kind of thing, right? And so I kind of hate it when innocent Chinese people abroad or whatever are just like attacked on this basis or what have you. Plus, the other thing is that many Chinese people say, well, look, relative to, you know, where we were when Deng took over in 1978, we built up the entire country. We're not starving to death anymore and the West wants to recolonize us. And so I understand where that's coming from. This way you want to be able to argue different points of view. With that said, there's one huge difference, right? Which is Nazi Germany was like 70 million people and the US was 150 million and the Soviet Union was 150 million and the UK was like 50 million. So they were outnumbered like five to one. </s><s>[INST] China outnumbers the US four to one. [/INST] This is going to be a fun century under this model. Under this model, things are going to be potentially crazy. Plus, you know, people are like, oh, I think this is, you know, again, I have nothing personal, you know, there's this guy Peter Zihan, he writes these books, right? I probably agree with about 20 or 30%, but I disagree with a lot of the rest. And a bunch of it is basically about how China's really weak and America's really strong and the rest of the world is screwed. And, you know, I think there's absolutely problems in China and, you know, like the current management is actually messing a lot of things up. We could talk about that. But I do think that, you know, the US is like fighting its factory. So one thing, you know, Zihan will talk about is how, oh, America has this blue water Navy, all the aircraft carriers in China has nothing, it's got bookcase, et cetera. Well, China ships things all around the world, right? It probably has, you know, one of the most active fleets out there in terms of, you know, it's commercial shipping. And in terms of building ships, here's a quote, China's merchant ship building industry is the world's largest, building more than 23 million gross tons of shipping in 2020. US yards built a mere 70,000 tons the same year that they typically average somewhere in the 200,000. That is a 100 to 300 X ratio, just in ship building. Pretty much everything else you can find in the physical world is like that, okay? We're not talking like 2X. We're talking, they can put together a subway station in nine hours with prefab and the US takes three years, okay? When you have a thousand X difference in the physical world, the reason the US was one against, you know, Nazi Germany in a serious fight is they had this giant manufacturing plant that was overseas and they just out produced, right? And they supplied the Soviets also with lendlies. And the Soviets talked about how they would not have won the war without the Americans. People are like, oh, the Russians, you know, fought the Germans. The Russians armed by Americans fought the Germans. Like it's a Soviet Union. They're not actually able to make high quality stuff. There obviously are individual people in the Soviet Russia who were innovative, right? I'm not taking that away. There's a tradition of amazing. I just want to be like, there's individual Russians who, obviously I admire Mendeleev and, you know, Kamo Groves and so on. There's amazing Russian scientists in here. </s><s>[INST] So I'm not saying that- I mean, in general, from building folks like yourself that criticize communism, it's too easy to say nothing communism produces is good, which of course is not true. Yes. A lot of brilliant people in, and then a lot of even, you know, there's a lot of amazing things [/INST] that have been created. A lot of brilliant people. Yeah. So they had some amazing mathematicians, amazing scientists and so on, right? </s><s>[INST] However- Great branding on the, you know, red and yellow. Yeah. The branding is stellar. Nazi Germany too, excellent branding. With the flag and so on, you know, yeah. [/INST] So- And ends there in terms of compliments. Yeah, yeah. Well, actually they copied a lot of stuff from each other. You know, like there's this movie called the Soviet Story. It basically shows a lot of Nazi and Soviet propaganda things next to each other. And you can see guys almost in like the same pose. It's almost like, you know how AI will do like style transfer? You can almost see, because the socialist realism style of like the muscular brawny worker, very similar to like the style of the Aryan Superman, you know, like pointing at the vermin or whatever. </s> |
<s>[INST] the love, but some applied social tech align, align with love, align with love. Because I have a bunch of really dumb ideas I'd prefer to hear less. I'd like to hear some of them first. I think the idea I would have is to be a bit more rigorous in trying to measure the amount of love you add or subtract from the world in second, third, fourth, fifth order effects. It's actually, I think, especially in the world of tech, quite doable. You just might not like, the shareholders may not like that kind of metric, but it's pretty easy to measure. That's not even, I'm perhaps half joking about love, but we could talk about just happiness and well-being, long-term well-being. It's pretty easy for Facebook, for YouTube, for all these companies to measure that. They do a lot of kinds of surveys, they could do, I mean, there's very simple solutions here that you could just survey how, I mean, surveys are in some sense useless because they're a subset of the population, you're just trying to get a sense, it's very loose kind of understanding, but integrated deeply as part of the technology. Most of our tech is recommender systems, most of the, sorry, not tech, online interactions driven by recommender systems that learn very little data about you and use that data, mostly based on traces of your previous behavior to suggest future things. This is how Twitter, this is how Facebook works, this is how AdSense for Google, AdSense works is how Netflix, YouTube work and so on. For them to just track as opposed to engagement, how much you spend on a particular video, a particular site, is also to track, give you the technology to do self-report of what makes you feel good, of what makes you grow as a person, of what makes you the best version of yourself, the Rogan idea of the hero of your own movie and just add that little bit of information. If you have people, you have this like happiness surveys of how you feel about the last five days, how would you report on your experience? You can lay out the set of videos, it's kind of fascinating to work, I don't know if you ever look at YouTube, the history of videos you've looked at, it's fascinating, it's very embarrassing for me. Like, it'll be like a lecture and then like a set of videos that I don't want anyone to know about, which will be like, I don't know, maybe like five videos in a row where it looks like I watched the whole thing, which I probably did, about like how to cook a steak even though or just like with the best chefs in the world cooking steaks and I'm just like sitting there watching it for no purpose whatsoever, wasting away my life or like funny cat videos or like legit, that's always a good one and I could look back and rate which videos made me a better person and not. And I mean, on a more serious note, there's a bunch of conversations, podcasts or lectures I've watched, which made me a better person and some of them made me a worse person. Quite honestly, not for stupid reasons, like I feel dumber, but because I do have a sense that that started me on a path of not being kind to other people. For example, I'll give you for my own, and I'm sorry for ranting, but maybe there's some usefulness to this kind of exploration of self. When I focus on creating, on programming, on science, I become a much deeper thinker and a kinder person to others. When I listen to too many, a little bit is good, but too many podcasts or videos about how our world is melting down or criticizing ridiculous people, the worst of the quote unquote woke, for example, all there's all these groups that are misbehaving in fascinating ways because they've been corrupted by power. The more I watch criticism of them, the worse I become. And I'm aware of this, but I'm also aware that for some reason, it's pleasant to watch those sometimes. And so for me to be able to self report that to the YouTube algorithm, to the systems around me, and they ultimately try to optimize to make me the best version of myself, which I personally believe would make YouTube a lot more money because I'd be much more willing to spend time on YouTube and give YouTube a lot more of my money, that's great for business and great for humanity because it'll make me a kinder person. It'll increase the love quotient, the love metric, and it'll make them a lot of money. I feel like everything is aligned. And so you should do that, not just for YouTube algorithm, but also for military strategy and for whether to go to war or not, because one externality you can think of about going to war, which I think we talked about offline, is we often go to war with kind of governments with, not with the people. You have to think about the kids of countries that see a soldier. And because of what they experience, their interaction with the soldier, hate is born. When you're like eight years old, six years old, you lose your dad, you lose your mom, you lose a friend, somebody close to you, the one really powerful externality that could be reduced to love, positive and negative, is the hate that's born when you make decisions. And that's going to take fruition, that that little seed is going to become a tree that then leads to the kind of destruction that we talk about. So but in my sense, it's possible to reduce everything to a measure of how much love does this add to the world. All that to say, do you have ideas of how we practically build systems that create a resilient society? [/INST] There are a lot of good things that you shared where there's like 15 different ways that we could enter this that are all interesting. </s> |
<s>[INST] Thanks for listening to this conversation with Constantine Batygin and thank you to squarespace little Roddy on it and and I check them out in the description to support this podcast. And now let me leave you with some words from Douglas Adams in the Hitchhiker's Guide to the galaxy Far out in the uncharted backwaters of the unfashionable end of the western spiral arm of the galaxy lies a small un regarded yellow sun Orbiting this at a distance of roughly 92 million miles is an utterly insignificant little blue green planet whose ape descendant life forms are so amazingly primitive that they still think digital watches are a pretty neat idea. Thank you for listening. I hope to see you next time. [/INST] </s> |
<s>[INST] Can I ask you a difficult question. So there's this famous um photograph of a falling man, so many people had to decide when they're above the fire in the fire whether to jump out of the building or to burn to death. What do you make of that decision? What do you make of that situation. [/INST] Those people who jumped those were acts of sheer desperation. I've been in fires and just minor burns but minor, you know, in situation, but I've been trapped or caught somewhat ended up in the burn center for some nothing, nothing serious at all. But like, but I I for those brief seconds, half a minute was, thank God if I didn't have my fire gear on, I would have been burned to very, very horrible level. Those people we're burning alive and they had the choice of either to stay there and burn alive or two launch themselves. And some of them, I don't fault them, but they had a few folks, they won't show it anymore because they say, I don't know why I defend some people, but they had a couple of folks that took umbrellas and they took garbage bags because they thought that it would slow down there except the acceleration rate to the ground and maybe just maybe they wouldn't be killed. And that's to me, a true sense of desperation for humanity to say I'm going to die either way. But let me take my chance and I don't know the exact number of those folks who did that, but our first Member of the Fire Department killed firefighter Daniel sir from 8 to 16 was struck by a jumper and one of my dear friends was ordered to help take him and they knew he was passed away because he was hit by a Flying Missile. I mean 120 mph body lands on you. Those those two bodies are now crushed And they were ordered to take that firefighter and bring him across the street to engine 10 Ladder 10. It was literally a firehouse less than 100 yards from the facade of the trade center from the trade center complex, they literally right there and there was plane parts that went into that firehouse landed into the front doors onto the roof. But the building itself was not destroyed. So it was used as a mini command center for quite a while. So my friend was ordered to take Daniel's body in respect and bring it over to this firehouse and give it some semblance of dignity and lay it out on one of the the bunk room, the bunks we have in the bunkhouse and just cover it with a sheet and put a sign please firefighter kill do not disturb and then we'll get to him later because obviously this operation is going to go on for days and my friend who is such a great, great, wonderful guy is so still to this day filled with guilt because if they weren't taking his body out with the respect and dignity that they did, it took a while because you know it's just it's a tough situation, his ladder company, it's coming over the bridge, there's a famous picture Vlad in 118, you see this tractor trailer fire truck is the one with the guy in the back also drives and it's a zoomed out shot and you see the Brooklyn bridge and you see only the fire truck in the middle and you see the two burning towers in the distance. Well his engine company was just ahead of them on the bridge and the only reason that engine company lived is their initial duty assignment was to take that firefighter and bring his body over. It's like the military, we don't leave anyone behind. These are our guys as we, some guys say it's all about the guy right next to you and nothing else really matters when that guy right next to you goes down, it stops, you get that guy to safety or if he's dead, you get him out. So in that time frame that saved his life. But that's a heavy burden to carry now for the rest of your life because you say if I wasn't helping my dead friend, I'm dead. Yeah. </s><s>[INST] What did it look like a ground zero? What did it feel like? What did it smell like? What you said it? There is a sense that it was almost like a war zone. But can you paint the picture of um how much dust is in the air? How hot is it? How many people are there? And again, how did it feel like [/INST] it was just um, it was a scene of controlled chaos control because there was a semblance of command and we were just trying to do our jobs, but it was such a frantic pace because we're now digging frantically knowing that there's life underneath this pile </s><s>[INST] and this is throughout the afternoon of [/INST] that. Yeah, I mean this was nonstop, you know, uh just nonstop really for for days. But from my particular crew, we literally kept going. We initially were dispatched over towards number seven, had </s><s>[INST] just gone down, [/INST] We were searching the post office that was there, there was reports of people trapped and we painstakingly searched every single inch of that building to make sure no one was left in there. And then we were deployed to the pile and the piles sort of ambiguous because it was just such a vast, vast pile. I mean, it went for city blocks and we were we were assisting in the retrieval of two port Authority police officers were lucky enough to survive, but they were trapped, they were deep down into a crevasse and they had to be physically dugout and extricated. So there was a couple of 100 few 100 guys involved in that process of bringing in equipment, jaws of life airbags to lift steel, you just to cut pieces of steel. It was just a huge operation and we were back toward the logistics end of it. Shuttling in gear and and bring it, bring in stretchers, bringing in oxygen, you know, whatever whatever was needed. And you were trying to climb over this, this jagged pile of debris. It wasn't like you just walked 100 ft on a on a street with something, you were trying to climb over this I beam and then down into this hole and then I'll back up that whole. I mean, just to run one piece of equipment, took a half an hour to get 100 ft, 200 ft mind you, some of these pieces of equipment, £100 you know, generator for her stools, this massive motor on a frame, </s> |
<s>[INST] Is there some at the highest level plays? There's still a role for feel [/INST] absolutely if you're not doing that, because here's the thing, right? No human being plays perfectly balanced in game three often like a robot. Would there not there not? Right. So there are opportunities there to take advantage of the things that they do that are slightly too aggressive or less aggressive, you know, for example, say most human beings don't bluff enough in a certain spot. So you don't have to call with the correct range of hands. You don't have to because they're not bluffing at the optimal frequency. So you don't have to call the optimal frequency. You'd be making this mistake, frankly, if you did </s><s>[INST] what's the difference between in person and online play given that context? [/INST] Yeah. Well, online poker and live poker, it's the same game, right? Same. It's poker, but it's different in so many levels. Right? Um, I think playing online, you have to focus far more on fundamentals, you know, on game theory, you don't have the added bonus of looking across the table and getting any sense of whether your opponent is strong or weak, they're bluffing whatever, you know? And and also Because online poker, those that play it, you play far more hands. Like some of these guys are playing 10, 20 tables at the same time, right? So you're just you're hitting the long run really quickly and you're creating a database on your opponent's, right? So let's have you know online I can see your data. Like this guy, he's playing 40% of hands, he's betting the river 80% of the time. So now I can use that data and you know, explain you that way when you play live you don't have that. </s><s>[INST] Do you enjoy playing online? [/INST] I enjoy, so with online poker I enjoy the convenience of it because you know, you can be on your couch in your underwear, not leave your house. Do </s><s>[INST] You also play multiple games at the same time? We do try to play one game. [/INST] I typically like to play one or two, but I can play up to four. I find that past four it's hard for me to keep up and keep track of what's actually happening. You know, it's a different mindset required. Like a lot of these young guys, they're accustomed to 20 tables at a time. It feels </s><s>[INST] like the purity of the game is gone. It's [/INST] Much more robotic, right? So if you're playing 20 tables, you're just making decisions based on like what, you know, you're not thinking about the the depth of the situation and what just happened 15 minutes ago, you don't even know what happened because you can't pay attention to all that at once. </s><s>[INST] And some of the magic of poker is the little sample [/INST] like for </s><s>[INST] example, I'm sorry to be bringing up magnets so much, but there's so much parallel between the two you and the poker and the chess world. He hates olympics and World Championships and all that kind of stuff because it's so little sample. But to me that's part of the magic of it. There's the World Series of poker, the main event, there's a magic to it [/INST] and </s><s>[INST] I don't know what that is exactly, because so much is at stake is so rare, so much drama and heartbreak leading up to it at all somehow. Um, yeah, it's, it accumulates to that magical moment when somebody wins, [/INST] especially that event, the World series of Poker main event historically. Like that's it, you know, that's the pinnacle, That's where like mainstream watches, that's where people are tuning in and the gravity of the moment. You know, it's so much bigger than people. Like everyone gets the opportunity to play armchair quarterback too, right? Oh, he should do this. You're not there, you're not under the lights, you're not under the pressure. You know, it might seem easy for you at home to be like, well I have because you can see the whole cards, you know, they can't certainly the idea of the small sample of tournaments, I like the idea that you don't have to worry about, oh, well if I do this now, then in the future, you know, I won't be balanced. I have to be balanced here or anything like that, that's like really boring and lame, right? Again, That is kind of the way the younger generation learns how to play the game being balanced in every spot. Um, and then randomizing, you know, like, oh, I'm supposed to do this 50% of time. Okay, so if my left card is red, I'll do it and if it's black, I don't, so you're not even making, you're no longer making actual decisions for yourself. You're just randomizing and that's way less fun for me then tailoring it to the situation </s><s>[INST] and the final table at the main event, There's none of that. You have to, it's all or nothing. Well, [/INST] you shouldn't be. But there are like, again, I think a lot of the young guys, they are thinking in that regard, like, oh, randomization </s><s>[INST] Maybe at that table, the final table at the main event, uh, what's a hand that stands out to you? That was especially gutsy and powerful or memorable for that you've seen in the history of poker? [/INST] Well, for me, the one that stands out and probably because I was so young and it was my first year, like one of, well, I want a bracelet that year. I was, I was friends with scotty win, the prince of Poker and he was heads up against a guy named kevin Mcbride and I was on the rail. You know, I'm like, wow, he's gonna, you know, his head's up and he was so cool. like he had a mullet, but it's perfect, right? He had a white shirt, the black thing, he's drinking a Michelob, smoking a cigarette, whatever, you know, all chill, he bets it all on the river and the guys thinking, and he psychologically owned him and he said, he goes with his beard goes, you call gonna be all over baby, that's right, okay. So this guy who was an amateur heard that and it was like, there's so much pressure in this moment right now, I can't handle this pressure, but scotty just told me if I call here, it's the pressure is gone, I don't have to be under it anymore. So he sort of hypnotized him into making the call, you know, and scotty had it scotty had, you know, the full house and it was over for the guy you call gonna be all over baby. It was, I just, I love that aspect, sort of the table talk dynamic, which isn't as prevalent today as it was back then. Um, but that one sticks out and it probably because it was one of my first, </s> |
<s>[INST] So that's really interesting. So I guess the presence of compute and the President's supervised data allowed the empirical evidence to do the convincing of the majority of the computer science community. So I guess there's a key moment with Jitendra Malik and uh Alex Alyosha Effros, who were very skeptical, right? And then there's a Geoffrey Hinton that was the opposite of skeptical and there was a convincing moment and I think emission had served as that moment and that represented this kind of where the big pillars of computer vision community, kind of, the wizards got together and then all of a sudden there was a shift and it's not enough for the ideas that I'll be there and the computer be there. It's for it to convince the cynicism that existed that that's interesting that people just didn't believe for a couple of decades. [/INST] Yeah, Well, but it's more than that, it's kind of when put this way, it sounds like, well, you know, those silly people who didn't believe what we're what we're missing. But in reality things were confusing because neuron really did not work on anything and they were not the best method on pretty much anything as well. And it was pretty rational to say, yeah, this stuff doesn't have any traction. And that's why you need to have these very hard tasks which are which produced undeniable evidence and that's how we make progress and that's why the field is making progress today because we have these hard benchmarks which represent true progress and so and this is why we were able to avoid endless debate. </s><s>[INST] So incredibly you've contributed some of the biggest recent ideas in ai in computer vision language, natural language processing, reinforcement learning sort of everything in between? Maybe not Ganz is there? There may not be a topic you haven't touched. And of course, the fundamental science of deep learning. What is the difference to you between vision language and as in reinforcement learning action as learning problems. And what are the commonalities do you see them as all interconnected? Are they fundamentally different domains that require different [/INST] approaches? Okay, that's a good question. Machine learning is a field with a lot of unity. A huge amount of unity. </s><s>[INST] What do you mean by unity? Like overlap of ideas [/INST] overlap of ideas, overlap of principles. In fact, there is only one or two or three principles which are very, very simple and then they apply in almost the same way in almost the same way to the different modalities to the different problems. And that's why today when someone writes a paper on improvement, optimization of deep learning envision it improves the different NLP applications and it improves the different reinforcement learning applications, reinforcement learning. So I would say that computer vision and NLP are very similar to each other today. They differ in that they have slightly different architectures, we use transformers in NLP and use convolutional neural networks envision, but it's also possible that one day this will change, everything will be unified with a single architecture because if you go back a few years ago in natural language processing, there were a huge number of architecture for every different tiny problem had its own architecture today there's just one transformer for all those different tasks and if you go back in time even more you had even more and more fragmentation and every little problem in ai had its own little sub specialization and sub little set of collection of skills, people who would know how to engineer the features. Now it's all been subsumed by deplaning, we have this unification and so I expect vision to become unified with natural language as well or rather I didn't expect, I think it's possible. I don't want to be too sure because I think on the commercial, you know that is very competition. The efficient a rail is different. A real doesn't require slightly different techniques because you really do need to take action. You really do need to do something about exploration, your variance is much higher. But I think there's a lot of unity even there and I would expect for example that at some point there will be some broader unification between RL and supervised learning where somehow they are l will be making decisions to make the super fast, don't even go better and it will be I imagine one big black box and you just throw you know, you shovel shovel things into it and it just figures out what to do with it, whatever you shovel at it. </s><s>[INST] I mean reinforcement learning has some aspects of language and vision combined almost there's elements of a long term memory that you should be utilizing and there's elements of a really rich sensory space. So it seems like the It's like the Union of the two or something like that, [/INST] I'd say something slightly differently. I'd say that reinforcement learning is neither, but it naturally interfaces and integrates with the two of them. </s><s>[INST] You think action is fundamentally different. So yeah, what is interesting about what is unique about policy of learning to act [/INST] Also, one example, for instance is that when you learn to act your fundamentally in a non stationary world because as your actions change the things you see, start changing you you experience the world in a different way and this is not the case for the more traditional static problem, you have some distribution and just to play a model to that distribution. Mhm. </s><s>[INST] You think it's a fundamentally different problem or is it just a more difficult generally it's a generalization of the problem of understanding. [/INST] I mean it's it's it's a question of definitions, almost there is a huge um you know, there's a huge amount of commonality for sure. You take gradients, you try, you take gradients, we try to approximate gradients in both cases and some in the case of reinforcement learning, you have some tools to reduce the variance of the gradients, you do that, there's lots of commonalities, the same neural net in both cases you compute the gradients, you apply adam in both cases, so I mean there's lots in common for sure, but there are some small differences which are not completely insignificant, it's really just a matter of your your point of view, what frame of reference, how much do you want to zoom in or out as you look at these problems, </s> |
<s>[INST] Do you have a word you prefer? Because we said faxes in the right word is their word like concepts? [/INST] I I would say assertions the assertions or rules because I'm not talking about rigid rules, but rules of thumb. But assertions um, is a nice one that covers um, all of these things. </s><s>[INST] Yeah. As a programmer to me assert has a very dogmatic authoritarian field to them. I'm so [/INST] sorry. Okay. But </s><s>[INST] assertions works okay. So if we go back to that room when Marvin Minsky with you all these seminal figures, uh, uh, Ed Fagan Mom, uh, thinking about this very philosophical, but also engineering question. We can also go back a couple of decades before then and thinking about artificial intelligence broadly when people were thinking about, you know, how do you create super intelligent systems? General intelligence? And I think people's intuition was off at the time. And I mean, this continues to be the case that we're not we're grappling with these exceptionally difficult ideas were not always, it's very difficult to truly understand ourselves when we're thinking about the human mind to to introspect how difficult it is to engineer intelligence to solve intelligence. We're not very good at estimating that. And you are somebody who has really stayed with this question for decades, Do you? What's your sense from the 1984 today? Have you gotten a stronger sense of just how much knowledge is required? So you've kind of said with some level of certainty that's still on the order of magnitude of tens of millions, but [/INST] For the first several years I would have said that it was on the order of one or two million. And so, um, it took, it took us about five or six years to realize that we were off by by a factor of 10. </s><s>[INST] But I guess what I'm asking, you know, Marvin Minsky is very confident in the 60s when you're right, what's your sense? If you, You know, 200 years from now, you're still, you know, you're, you're, you're not going to be any longer in this particular biological body, but your brain will still be, uh, in the digital form and you'll be looking back, would you think you were smart today? Like your intuition was right? Or do you think you may be really off? [/INST] So I think I'm, I'm right enough and let me explain what I mean by that, which is, um, sometimes like if you have an old fashioned pump, you have to prime the pump and then eventually it starts. So I think I'm, I'm right enough in the sense that what we, what we've built, even if it isn't, so to speak, everything you need. Um, it's primed the knowledge pump enough that psych can now itself help to learn more and more automatically on its own by reading things and understanding and occasionally asking questions like a student would or something and by doing experiments and discovering things on its own and so on. So through a combination of psych powered discovery and psych powered reading, it will be able to bootstrap itself, Maybe it's the final 2%, maybe it's the final 99%. So even if, I'm if I'm wrong, all I really need to to build is a system which has primed the pump enough that it can begin um that cascade upward, that self reinforcing sort of quadratic lee or maybe even exponentially increasing um um path upward um that we get from for instance, talking with each other, that's why um humans today know so much more than humans 100,000 years ago. We're not really that much smarter than people were 100,000 years ago, but there's so much more knowledge and we have language and we can communicate um we can check things on google and so on. So effectively we have this enormous power at our fingertips um and there's almost no limit to how much you could learn if you wanted to because you've already gotten to a certain level of understanding of the world that enables you to read all these articles and understand them, that enables you to go out and if necessary, do experiments, although that's slower as a way of gathering data um and so on and I think this is really an important point, which is if we have artificial intelligence real general artificial intelligence human level artificial intelligence then people will become smarter. Um It's not so much that it will be us versus the ai is it's more like us and the A. I. S. Together, we'll be able to do things that require more creativity. Um That would take too long right now. But we'll be able to do lots of things in parallel. We'll be able to misunderstand each other less. Um uh There's all sorts of value that effectively for an individual would mean that individual will for all intents and purposes be smarter. And that means that humanity as a species will be smarter. And when was the last time that any invention qualitatively imp made a huge difference in human intelligence. Um You have to go back a long ways. It wasn't like the internet or the computer or mathematics or something. It was all the way back to the development of language. We sort of look back on pre linguistic caveman as well, you know, they weren't really intelligent were they weren't really human were they? And I think that um as you said, 510,200 years from now, people will look back on people today um right before the advent of these sort of lifelong general ai muses. Um and say you know those poor those poor people, they weren't really human, were </s> |
<s>[INST] is the ability to constantly do refoundings rebirths. So if you imagine a world that is primarily network states, can you help me imagine what that looks like? Now, there's several ways to imagine things, which is how many of them are there? And how often do they new ones pop up? There could be 1000s, given 7 billion [/INST] people, 8 billion people on earth. Yeah, so there's network state in the, like the precise definition I have in the book, which is a diplomatically recognized entity. And there's never seen sort of the loose definition where you know, one thing that's interesting is this term has become a lowercase term really fast. Okay, no, you stay. Yeah. Like in the sense of Google, became lowercase Google for like googling or like Uber became lowercase Uber. Like if you go to the network state comm front, such reviews, or you go to search when search.tv.com and put in their network state you see it's become like a word orily phrase. So that means whatever I intended to be the people will use it to mean they want to be. Right? Internet. </s><s>[INST] It's internet, right? What is your meme, and what is your book? Am I a meme? [/INST] Ok maybe I'm a meme. Can I put that into a good meme? I wanted people to take it out there and make it their own. One of the things I say at the beginning and I go back to this thing it's a toolbox not a manifest. Even if you dislike 70% of it, 80% of it, 90% of it, if there's something that's useful to you, you can take that and use it just like a library, a software library. You might just use one function there. Great. I'm glad. </s><s>[INST] I've delivered you some value. That's my purpose in this. [/INST] That's my purpose in this. So you're not Ayn Rand. No, I'm not Ayn Rand. Basically, the whole point of this actually is it's polytheistic, polystatistic, polynemistic is genuinely- Is it polyamorous? It's not polyamorous. Okay. Though somebody might- </s><s>[INST] Do you have love- Somebody might- Do you have love advice in the book? I didn't see it. So did you talk about love in- I do not talk about love. [/INST] No one say it. I do not talk about love. I would rather- Maybe not to you. </s><s>[INST] Not that I don't believe in love. Love is great. Maybe not to you. Not that I- All right. [/INST] I will accept your offer to write a guest chapter in your V2 book about love. </s><s>[INST] All right. Great. Because there is some aspect that's very interesting which parts of human civilization require physical contact, physical, because it seems like more and more can be done in a digital space. Yeah. But as I said, we work. [/INST] Yeah, but, uh, as I said, we work, for example, but you're not going to build a self-driving car city in digital space. You're not going to be able to do cars at all. Well, you're, well, so sure. But let's say you're not going to be able to get to Mars in a purely digital thing you need to build. You know, you have a little rocket launch pad, you're not going to be able to do all the innovative by medicine, whether it's, you know, all the. If you've seen bioelectricity or there's stuff on regenerative medicine, all this stuff, you just can't do that digitally, right? We're still physical beings, so you need physical space, but how do we get that, right? This is meant to wend its way through various roadblocks in the so-called, actually my term from many years ago, the idea maze, it's meant to wend its way through the idea maze to find how to use bits to re-unlock innovation in atoms. </s><s>[INST] The idea maze within the bigger prime number maze, or go back to visualizing the number [/INST] of states and how often are they born, or go back and how often are they born. So let me first anchor this because people, just to give some numbers, right? How many UN-listed countries are there? Like 196, 193, okay? And there's some that are on the border like Taiwan or Israel, right, where they're not, I mean, Israel is a country, but it's not recognized by every country or what have you, right? Is Texas a country? No, but it may eventually become, right, okay. So within that list of about 200 countries, okay? I've got a graph in the book that shows that most countries are actually small countries. There's 12 countries that have less than a 100,000 people by the UN definition of a country. </s><s>[INST] There's another 20 something that have between a 100,000 and one million, may eventually become. [/INST] There's another 50 or 60 something that have between a million and 10 million. So most countries in the UN are less than 10 million people. There's only 14 countries that are over 100 million people, okay. So most countries are small countries is kind of surprising to us because most people live in big countries. So now you're like okay well, I've built social networks that are bigger than that. You have a following that's bigger than 100,000 people. You have a following that's bigger than, small country like Kiribati or whatever, right? So that first changes feasibility. You think of a country as this huge thing but it's actually smaller than many, many countries are smaller than social networks that you've built. Number one. is the number of UN listed countries, even though it's being flat-ish for the last 30 years, with like a few things like South Sudan and East Timor that have come online, there's a graph that I posted which shows that it's increased by about 40 or 50 something at the end of World War II, when the UN was set up to 197 today. There's been like kind of a steady increase in particular with all the decolonization, all the countries that got there, and it depends first from the British Empire, and then from the Soviet Empire, right? That imperial breakup led to new countries, okay? And so then the question is, is that flat forever? Well, the number of new currencies similarly increased for a while, roughly one per country or thereabouts, and then it was flat for a while, and then suddenly it's gone completely vertical. That's an interesting graph, right? Where it's like linear-ish, then it's flat, and then it just goes, whoof, like this. Now you can define, you can argue where the boundary is for a new currency, okay? But I think Bitcoin certainly counts, I think Ethereum certainly counts in terms of just its scale and adoption worldwide. So at least you have two. If you take the Broadchurch view, you have a thousand or something like that, right? Somewhere in between, you might say, how many currencies are above the market cap of an existing previously recognized fiat currency, like which got onto the leaderboard, right? There's a website just like coinmarketcap.com, that's like a site for like cryptocurrency tracking, it's very popular, okay? There's a fun site called fiatmarketcap.com which shows where Bitcoin is relative to the fiat currencies of the world. And it's like last I checked, number 27, somewhere in between the Chilean peso and the Turkish cholera or something, okay? And it previously being close to crack in the top 10, okay? And I think it will will again, at some point. So, we know that you can have a currency out of nowhere that ranks with the fiat currencies of the world, could you have a country out of nowhere that ranks with the countries of the world? So this is maybe the fastest way probably should have said this at the very beginning. If you go to the network state in one image, okay, that kind of summarizes what a network state looks like in a visual, just one single visual and the visual is of a dashboard. And the dashboard shows something that looks like a social network, except you're visualizing it on the map of the world. And it's got network nodes all over the place 100 people here 1000 people there, they're all connected together. The total population of the people in this social network is about 1 million people. So 1.7 million people in this example. And some of the buildings are some of the people are just singleton. So just folks in their apartment who can conceptualize themselves as citizens of this network state. And they've got the flag on their wall, right? And the digital passport on their on their phone along with digital currency. Others are groups of hundreds or 1000s or even 10s of 1000s of people that have all taken over a neighborhood just like Chinatowns exist, right? Just like, you know, intentional communities existed. They just basically, you know, go and crowdfund lane together, right? And these are all network together, you know, just like the islands of Indonesia are separated by ocean. These are islands of this network state that are separated by internet. Okay, so conceptualize themselves as something. And at the very top of the dashboard, there's something very important, which is the population annual income and real estate footprint of this network state. So population, we already discussed you can build an online social network, we know you can build something which has a population that's bigger than these 100,000 or million person countries. One of the new things contributions a network it has is say that you cannot just exceed in population you can exceed in real estate footprint. Because one way of thinking about it is, on I don't exactly know the numbers on foreign ownership in Estonia, but let's say to First Order the million-something Estonians own and could afford Estonia. Okay, a million people could buy a territory that is the size of Estonia, right? That's probably true to First Order. There might be some overseas ownership, but it's probably true. You probably find a country for which that's true. What that means is, a million people Dhisialee could buy distributed territory that is probably greater than or equal to the size of Estonia. Especially if they're buying desert territory or stuff like that, which means now you have a digital country that is ranking, not just in people, not just in real estate footprint, you know, so it's also in real estate footprint with the with the with the countries of the world. So you start ranking, and you're bigger than these UN listed countries in your population, and your real estate footprint. And the third is income. Okay. You can prove on chain, that you have a income for the for the digital population that is above a certain amount, right? This is what I call the census of the network state. And it's actually such a crucial component that I I have it in the essay, The New York State in a Thousand Words, the post office and census were actually important enough to be written into the US Constitution, partly because it was for apportionment of representatives, partly because there's a feedback mechanism. That census was done every 10 years and provided a crucial snapshot of the US for the last several hundred years. Now, here, this census of a digital state could be done every 10 seconds. Conducting it is actually not the hard part. The hard part is proving it, because how will the world believe that you actually have 100,000 people spread across countries? Couldn't there all be bots? Could there be AIs? Proof of human, proof of income, and also proof of real estate start to actually rise dramatically in importance because you're saying we're going to rank this digital state on the leaderboard of the fiat states. That means that people will start to... At first, they'll just laugh at it. Once you start claiming you have 10,000 citizens, people are going to start poking and be like, is that real? Prove that it's real. I have a whole talk on this, actually, I'm giving at this Chainlink conference, but essentially, how do you prove this? The short answer is crypto oracles plus auditing. The somewhat longer answer is you put these assertions on chain, these proof of human, these proof of real estate, et cetera, assertions on chain. There's people who are writing to the blockchain and they are digitally signing their assertions. Now, of course, simply just putting something on chain doesn't make it true. It just says, you can prove not that what is written on chain is true but that the metadata is true. You can show who wrote it via their digital signature, what they wrote, their hash, and when they wrote their timestamp. You can establish those things in metadata, who, what, and when was written. Who is the who in that picture? </s> |
<s>[INST] and that event the fatality that's connected to an autonomous vehicle is significantly at least currently magnified in the amount of uh attention again. So that speaks to public perception. I think the most popular topic about autonomous vehicles in the public is um the trolley problem formulation right? Which has let's not get into that too much but is misguided in many ways. But it speaks to the fact that people are grappling with this idea of giving control over to a machine. So how do you win the the hearts and minds of the people that autonomy is something that could be part of their lives, [/INST] how you let them experience it? Right. I think it's I think I think it's right. I think people should be skeptical. I think people should ask questions. I think they should doubt because this is something new and different. They haven't touched it yet and I think it's perfectly reasonable and but at the same time it's clear there's an opportunity to make the roads safer. It's clear that we can improve access to mobility. It's clear that we can reduce the cost of mobility and that once people try that and are understand that it's safe and are able to use in their daily lives. I think it's one of those things that will, we'll just be obvious. And I've seen this practically in demonstrations that I've given, where I've had people come in and you know, they're very skeptical and they get in a vehicle, you know, my favorite one is taking somebody out on the freeway and we're on the one the one driving at 65 miles an hour and after 10 minutes, they kind of turn and ask, is that all it does? And you're like, it's self driving car. I'm not sure exactly what you thought it would do, right. Um, but they, you know, they, it becomes mundane, which is, which is exactly what you want a technology like this to be right. We don't really, when I turned the light switch on and here, I don't think about the complexity of, you know, those electrons, you know, being pushed down a wire from wherever it was and being generated, like it's just, it's like I just get annoyed if it doesn't work, right? And and what I value is the fact that I can do other things in this space, I can, you know, see my colleagues, I can read stuff on a paper, I can, you know, not be afraid of the dark. And I think that's what we want this technology to be like, is it's it's in the background and people get to have those most life experiences and and do so safely? </s><s>[INST] So putting this technology in the hands of people speaks to scale of deployment. All right. So what do you think the dreaded question about the future? Because nobody can predict the future, but just maybe uh, speak poetically about when do you think we'll see a large scale deployment of Autonomous Vehicles? 10,000? Those kinds of numbers? [/INST] We'll see that within 10 years. I'm pretty confident. We um, </s><s>[INST] what's an impressive scale? What moment? Uh, so you you've done DARPA challenge with this one vehicle at which moment does it become? Wow, this is serious scale. [/INST] So, so I think the moment it gets serious is when we really do have driverless vehicle operating on public roads and that we can do that kind of continuously without </s><s>[INST] a safety without a safety [/INST] Driver in the vehicle. I think at that moment we've kind of crossed the 0-1 threshold and then it is about how do we continue to scale that? How do we build the right business models? How do we build the right customer experience around it so that it is actually, you know, a useful product out in the world. And I think that is really at that point it moves from a, you know, what is this kind of mixed science engineering project into engineering and commercialization and really starting to deliver on the value that we all see here. Uh, and you know, actually making that real in the world, what do you think </s><s>[INST] that deployment looks like? Where do we first see the inkling of No safety driver one or 2 cars here and there? Is it on the highway? Is it in specific routes in the urban environment? [/INST] I think it's gonna be urban, suburban type environments? Yeah, we with a roar when we thought about how to tackle this, it is kind of in vogue to think about trucking as opposed to urban driving. And, and, you know, the, again, the human intuition around this is that freeways are easier to drive on because everybody's kind of going in the same direction and, you know, lane's a little wider, et cetera. And I think that that intuition is pretty good, except we don't really care about most of the time. We, we care about all of the time. Uh, and when you're driving on a freeway with the truck, say, uh, 70, mph, Uh, you've got £70,000 load with you, that's just an incredible amount of kinetic energy. Uh, and so when that goes wrong, it goes really wrong and that, um, those, those challenges that you see occur more rarely, so you don't get to learn as as quickly. Uh, and they're, you know, incrementally more difficult than urban driving, but they're not easier than urban driving. And so I think this happens in moderate speed, Urban environments because they're, you know, if if two vehicles crashed at 25mph, it's not good, but probably everybody walks away and those, those events where there's the possibility for that occurring happened frequently. So we get to learn more rapidly, We get to do that with lower risk for everyone and then we can deliver value to people that they need to get from one place to another and then once we've got that solved, then the kind of the freeway driving part of this just falls out. But we were able to learn more safely more quickly in the urban environment. </s> |
<s>[INST] You can buy fractions of a stock, say $1 worth no matter what the stock prices, broker services are provided by cash up investing. A subsidiary of square and member S. I. P. C. I'm excited to be working with cash app to support one of my favorite organizations called first, best known for their first robotics and lego competitions. They educate and inspire hundreds of thousands of students in over 110 countries and have a perfect rating and charity navigator, which means that donated money is used to maximum effectiveness When you get cash app from the app store, Google play and use collects podcast. You'll get $10 in cash app will also donate $10 to first, which again, is an organization that I've personally seen, inspire girls and boys to dream of engineering a better world. And now here's my conversation with Grant Sanderson. [/INST] I think </s><s>[INST] there is intelligent life out there in the universe. Do you think their mathematics is different than [/INST] ours jumping right in? I I think it's probably very different. There's an obvious sense. The notation is different. Right? I think notation can guide what the math itself is. I think it has everything to do with the form of their existence. Right. </s><s>[INST] Do you think they have basic arithmetic? Sorry, interrupt. [/INST] Yeah. So I think they count. Right? I think notions like 123, the natural numbers. That's extremely well natural. That's almost why we put that name to it. Um, as soon as you can count, you have a notion of repetition, right? Because you can count by two two times or three times. And so you have this notion of repeating the idea of counting which brings you addition and multiplication. I think the way that we extend to the real numbers, there's a little bit of choice in that. So there's this funny number system called the surreal numbers that um, it captures the idea of continuity. It's a distinct mathematical object. You could very well model the universe in motion of planets with that as the back end of your math. Right? And you still have kind of the same interface with the front end of what physical laws you're trying to or what physical phenomena you're trying to describe with math. And I wonder if the little glimpses that we have of what choices you can make along the way based on what different mathematicians have brought to the table is just scratching the service surface of what the different possibilities are. If you have a completely different mode of thought or mode of interacting with the universe. </s><s>[INST] And you think notation is a key part of the journey that we've taken through math. [/INST] I think that's the most salient part that you've noticed. At first, I think the mode of thought is going to influence things more than like the notation itself, but notation actually carries a lot of weight when it comes to how we think about things more so than we usually give it credit for. I would I would be comfortable saying </s><s>[INST] you have a favorite or least favorite piece of notation in terms of its effectiveness. [/INST] Yeah. Well, so at least favorite one that I've been thinking a lot about that. We'll be a video. I don't know when but we'll see uh the number E. We write the function E to the X. This general exponential function with the notation. E to the X. That implies. You should think about a particular number, this constant of nature. And you repeatedly multiply it by itself. And then you say what's E to the square root of two? And you're like, oh well we've extended the idea of repeated multiplication. That's that's all nice. That's all nice and well. But very famously you have E to the pie. I like what we're extending the idea of repeated multiplication into the complex numbers. You can think about it that way. In reality I think that it's just the wrong way of um notation aly representing this function. The exponential function which itself could be represented a number of different ways you can think about it. In terms of the problem, it solves a certain very simple differential equation, which often yields way more insight than trying to twist the idea of repeated multiplication, take its arm and put it behind its back and throw it on the desk and be like you will apply to complex numbers, right? That's not I don't think that's pedagogical helpful. </s><s>[INST] And so the repeated multiplication is actually missing the main point. The power of E to the S. [/INST] I mean what what it addresses is things where the rate at which something changes depends on its own value. But more specifically, it depends on it linearly. So for example, if you have a population that's growing and the rate at which it grows depends on how many members of the population are already there. It looks like this nice exponential curve. It makes sense to talk about repeated multiplication because you say how much is there after one year, two years, three years you're multiplying by something. The relationship can be a little bit different sometimes. Where let's say you've got a ball on a string. Like a like a game of tether ball going around a rope, right? And you say it's velocity is always perpendicular to its position. That's another way of describing its rate of change is being related to where it is. But it's a different operation. You're not scaling it. It's a rotation. It's this 90° rotation. That's what the whole idea of like complex exponential nation is trying to capture. But it's obfuscated in the notation when what it's actually saying. Like if you really part something like eat the pie, I what it's saying is choose an origin. Always move perpendicular to the vector from that origin to you. Okay? Um Then when you walk pi times that radius, you'll be halfway around. Like that's what it's saying. Um It's kind of the U. Turn 90 degrees and you walk, you'll be going in a circle. That's the phenomenon that it's describing. But trying to twist the idea of repeatedly multiplying a constant into that. Like I can't even think of the number of human hours of like intelligent human hours that have been wasted. Trying to parse that to their own liking and desire among scientists or electrical engineers or students have. We were which if the notation, we're a little different or the way that this whole function was introduced from the get go were framed differently, I think could have been avoided. Right? </s> |
<s>[INST] And when you see those desperate lines all laid out on the page, how difficult is it to then start stitching it together? Do you find that the when you look at a list of those things, the final product will look very different or will you actually use those [/INST] lines? So I will use those lines then I have a file called scraps. So like if the line is no longer used, I put in my scrap pile, </s><s>[INST] I'd love to see what's in the scrap [/INST] pile. Okay. Yeah, sure. One of the things I've been pulling scraps is a lot of times when I was earlier writing, I would have contemporary references and I realized that that's bad because I want the The the reader to be in the past as the present. So if if you're talking about, let's say 1901 and then you're referring to Obama, that screws people up. So I had to pull all those. </s><s>[INST] Okay, let's talk about some New Year's resolutions. Did you ever do New Year's resolutions? You ever think like that? Like take a special day in the year to think about how you're going to try to change yourself? Or do you try to transform yourself every single day when you wake up? [/INST] Well, I usually have several projects, I'm working on it once. So there's always incremental progress on those, </s><s>[INST] you know, it's nice to have a deadline by the end of 2022 I'll accomplish this kind of uh like to hold yourself responsible and then you could do that at the beginning of the year to think about that both philosophically, like what kind of big not, not projects that you can quantify, but more like how can I change my life? Or like I mentioned, take the leap of different kinds and then there's specific things like finish the book [/INST] I years ago and I'm, I'm, I think on some level you much less than me, but I think you're increasing in this direction. I realized it's more, I have to learn how to be a, a surfer and not a driver because when you reach the level we're at in our careers or in our place in the culture, a lot of this is luck and a lot of this is just like, like I'm just going along for the ride because it's kind of counterintuitive, like, you know, like the success of the anarchist handbook was counterintuitive. Um so all I'm hoping for is, you know, getting the book done, I am extremely proud of it. Um and and just also, you know, building a, you know, we had thanksgiving together at Blair's house, just building a great upcoming community here in Austin, which is, it has happened very quickly, I was, there was going to be another surprise here, there's a girl named Natalie um side surf and she makes these ultra realistic cakes, like if you've seen those cakes online, what looks like cutting a puppy, like she makes those kind of things. So she's here um yeah, you know, so like, I think she's been here for a while. I've never, I haven't met her yet, but I just kind of chatted with her. So it's just, it's just so many, there's so many um scenes happening here um that are overlapping. </s><s>[INST] So in general finished the book. Keep building your community. I mean you've already been doing that here, you've been here several months. [/INST] I've been making a point to introduce people to each other and everyone's just really getting along very well. That's </s><s>[INST] great. And the book is the focus of the book [/INST] is the focus. </s><s>[INST] What about the podcast that you're doing? You're welcome. [/INST] Yeah, I, I mean I enjoy it and it's been growing a lot. I finally got a new computer which my friend jay installed, so I can have a decent camera because of my old um this is my mindset as a hoarder. Like I was more interested in spending money on a parade of autograph than actually getting a computer that's from the 20th century. Um so, but I, I'm such an old school person in that in my head podcasts are like, so ephemeral, like I don't like there's some episodes in my podcasts I'm really proud of and there's a lot of friendships I've made as a result of it, that really mean a lot to me, No question, it's made my life profoundly better place. But it's not the same as that book on the shelf, especially when the book is something that I think matters much more than I do. </s><s>[INST] Yeah. There's a permanence to it. There's a seriousness to laying down the words on paper, like really giving him thought. Yeah, that's true. But podcast, I mean, I'm I'm a huge fan of podcasts. You're just, you're just, you don't listen to podcasts much. It's just fascinating to [/INST] Yeah. Like at all. And I like, I don't know how mine is so successful. It's just Yeah. </s><s>[INST] Yeah. Yeah. I just love the medium. Yeah. But your, I love the authenticity, the realness of the medium. That's really nice. [/INST] I just understood for the it's starting to click because like my pal blair white, she was just done Rogan And the 1st 10 minutes I was, I was so angry. Like I was sitting there like yelling at the screen because Jo and Blair, you would think that they're going to start talking about you know, trump or trans issues or moving to Austin, They start talking about shark reproduction and neither these dumbasses knew anything about it and I know a lot about it and they're like, oh is it like this or do the sharks lay eggs. And I'm sitting there like, if you don't know what you're talking about this. Why, why are you talking? And I could also see why people like these shows because they feel like they're friends of the people. Like they're sitting in the room because I felt like it was in that room and I wanted to shake both of </s> |
<s>[INST] I try to make these interesting but I give you time stamps so you can skip but still please do check out the sponsors by clicking the links in the description. It's the best way to support this podcast. This show sponsored by Babbel, an app and website that gets you speaking in a new language within weeks go to babble dot com and use collects to get three months free. They offer 14 languages including Spanish, French, Italian German and yes, Russian Daily lessons are 10-15 minutes, super easy effective designed by over one [/INST] 100 language </s><s>[INST] experts. Let me read a few lines from the Russian poem notch ulitsa phonied up Jessica by alexander block that you'll start to understand if you sign up to babble notch bullets of an arab checa dismissing me to squeeze yet. Chevy show hot cheetos vodka shop talk is hotter yet. Now I say that you'll start to understand this poem because Russian starts with the language and ends with vodka. Now the latter part is definitely not endorsed or provided by Babel will probably lose me this sponsorship although it hasn't yet. But once you graduate with Babel you can roll my advanced course of late night Russian conversation over vodka, no out for that yet. So get started by visiting babble dot com and use culex to get three months free. This show is also sponsored by Master class, sign up at Masterclass dot com slash looks to get a discount and to support this podcast when I first heard about Master class, I thought it was too good to be true. I still think it's too good to be true for 100 $80 a year. You get an all access pass to watch courses from, to list some of my favorites chris Hatfield on space exploration, hope to have them in this podcast one day Neil Degrasse Tyson on scientific thinking, communication Neil too, will wright creator some city and sims and game design, Carlos, Santana on guitar, Garry Kasparov in chess, Daniel aground on poker and many more chris Hadfield explaining how rockets work and the experience of being watched at the space alone is worth the money by the way, you can watch it on basically any device once again, sign up Masterclass dot com slash likes to get a discount and to support this podcast. This show finally is presented by cash app, the number one finance app in the app store. When you get it used collects podcast cash app, lets you send money to friends buy Bitcoin and invest in the stock market with as little as $1 since cash, apology, send and receive money digitally. Let me mention a surprising fact related to physical money of all the currency in the world, Roughly 8% of it is actually physical money. The other 92% of the money only exists digitally and that's only going to increase. So again, if you get cash app from the app store, google play and use culex podcast, you get 10 bucks and cash app will also donate $10 to first an organization that is helping to advance robotics and stem education for young people around the world. And now here's my conversation with francois chalet thought philosophers, thinkers or ideas had a big impact [/INST] on you growing up and today. So one author that are the big impact on me. When I read his books as a teenager, I was jean pierre with swiss psychologist is considered to be the father of developmental psychology. And yes, a large body of work about um basically how intelligence develops in Children. And so it started all the work, like most of it is from the 1930s, So it's not quite up to date, it's actually supported by many neural developments in developmental psychology. But to me it was it was very, very interesting, very striking and actually shaped the early ways in which I started to think about the mind and the development of intelligence as a Tunisia. His </s><s>[INST] actual ideas or the way he thought about, or just the fact that you could think about the developing mind at [/INST] all. I guess both john pleasure is the author that introduced me to the notion that intelligence and the mind is something that you construct throughout your life and that you the Children constructed in stages. And I thought it was a very interesting idea which is of course very relevant to a i to building artificial minds. Another book that I read around the same time that had a big impact on me. Uh and and there was actually a little bit of overlap with the same period as well and I read it around the same time, is Jeff Hawkins on Intelligence, which is a classic and he has this vision of the mind as a multi scale hierarchy of tamper prediction modules. And this I just represented with me like the notion of uh murdering hierarchy um of you know, potentially um I've compression functions of prediction functions I thought was really interesting and it really shaped the way I started thinking about how to build minds </s><s>[INST] the hierarchical nature, the which aspect also, he's a neuroscientist and he was thinking actual, he was basically talking about how our mind works. [/INST] Had the notion that cognition is prediction was indicted, It was kind of new to me at the time and I really loved at the time. And yeah, and the notion that yet there are multiple scares of processing in the brain. The hierarchy. Yes, this is before deep learning these ideas of hierarchies in the area has been around for a long time, Even before our intelligence. I mean they've been around since the 1980s. Um and yeah, that was before the planning. But of course I think these ideas really found there practical implementation in the planning. </s> |
<s>[INST] What image selection mechanism would you say? Is there like uh inherent in some aspect of the process? So you're kind of implying it, there's almost none. But what what is there would you say if you introspect? [/INST] So it's not like a new curated can basically like one way of imagining and curated is basically you have cameras, cameras that can take pictures at random viewpoints. When people upload pictures to the internet, they are typically going to care about the framing of it. They're not going to upload, say the picture of a zoomed involved. For example, </s><s>[INST] when you say internet doing in social networks. Okay, [/INST] so these are not going to be like pictures of like a zoomed in the table or resumed involved. So it's not really completely amputated because people do have the like photographers bias where they do want to keep things towards the center a little bit or like really have like nice looking things and so on in the picture. So that's the kind of bias that typically exists in the state of it. And also the user base, right? You're not going to get lots of pictures from different parts of the world because there are certain parts of the world where people may not actually be uploading a lot of pictures on the internet or may not even have access to a lot of internet. </s><s>[INST] So this is a giant dataset and a giant neural network. I don't think we've talked about what architectures work well for ssl for self supervised learning [/INST] for sear. And first of all we were using convolutional networks, but recently in a word called Dino, we basically started using transformers for vision. Both seem to work really well uh continents and transformers and depending on what you want to do, you might choose to use a particular formulation. So for sure it was a continent. It was particularly a rig net model which was also work from facebook Rechnitz are like really good when it comes to compute versus like accuracy. So because it was a very efficient model, compute and memory vice efficient and basically it worked really well in terms of scaling. So we used a very large dragnet model and turn it on the building images. </s><s>[INST] You may be quickly comment on what rug nets are. Uh it comes from this paper designing network design spaces. It's just super interesting concept that emphasis on how to create efficient neural networks, large neural [/INST] networks. So, one of the key takeaways from this paper, which the authors like whenever you hear them present this work, they keep saying is lot of neural net books are characterized in terms of flops, threat flops basically being the floating point operations and people really love to use flops to say this model is like really computational heavy or like our model is competition achieved and so on. Now it turns out that flops are really not a good indicator of how well a particular network is, like how efficient it is really and what a better indicator is the activation of the memory that is being used by this particular model. And so designing like one of the key findings from this paper was basically that you need to design network families or neural network architectures that are actually very efficient in the memory space as well, not just in terms of pure flops. So right now it is basically a network architecture family that came out of this paper that is particularly good at both flops and the sort of memory required for it. And of course it builds upon like earlier work, like Resnet being like the sort of more popular inspiration for it where you have residual connections. But one of the things in this work is basically they also use like squeeze excitation blocks. So it's a lot of nice sort of technical innovation and all of this from prior work and a lot of the ingenuity of these particular authors in how to combine these multiple building blocks. But the key constraint was optimized for both flops and memory when you're basically doing this, don't just look at flops </s><s>[INST] and that allows you to what have sort of have very large networks through this process. Can optimize for low, like for efficiency [/INST] memory also. And just in terms of pure hardware, they fit very well on Gpu memory. So they can be really powerful neural network architecture with lots of parameters, lots of flops. But also because they're like efficient in terms of the amount of memory that they're using. You can actually fit a lot of these on, you can fit a very large model in a single Gpu for example, </s><s>[INST] would you say that the choice of architecture matters more than the choice of maybe data augmentation techniques? Is there a possibility to say what matters more? You kind of imply that you can probably go really far with just using basic continents? [/INST] All right. I think they like data and data augmentation. The algorithm being used for the self supervised training matters a lot more than the particular kind of architecture with different types of architecture. You will get different like properties and the resulting sort of representation. But really, I mean the secret sauce is in the data augmentation and the algorithm being used to train them. Um The architectures, I mean at this point a lot of them performed very similarly depending on the particular task that you care about. They have certain advantages and disadvantages. </s><s>[INST] Is there something interesting to be said about what it takes with Sears to train a giant neural network? You're talking about a huge amount of data. A huge neural network. Is there something interesting to be said of how to effectively train something like that fast? [/INST] Lots of gps. </s><s>[INST] Okay. [/INST] I mean, so the model was like a billion parameters uh and it was trained in the billion images. So like basically the same number of parameters as the number of images. And it took a while, I don't remember the exact number, it's in the paper, but it took a </s> |
<s>[INST] So that that seems like a seminal mistake by Harvard. And Harvard is one of the great universities in the world and so sort of its successes and its mistakes are really important for the world as a beacon of like how we make progress. So what lessons for the bigger academia that that's under fire a lot these days. Uh what bigger lessons do you take away? Like how do we make Harvard Great, How do we make other universities? Yale, M. I. T. Great in the face of such mistakes? [/INST] Well, I think that we have moved into a model where we the we have the consumerism nation of Education. Uh that is to say we have feckless administrators who make policy based on what the students say. Um Now this comment is not intended to suggest that students have no voice in governance, but it is just to suggest that the faculty are there for a reason. They are among the greatest minds on the planet earth in their particular fields, at schools like Harvard and Yale stanford the schools that you mentioned, m I. T. Quite literally the greatest minds on earth. Their therefore a reason, things like uh curriculum and so forth, are rightly in the province of faculty. And while you take input and critique and so forth, ultimately the grownups in the room have to be sufficiently responsible to take uh to to take charge and to direct the course of a student's education And um you know, you know, my situation is one example where it really could have been an excellent teaching moment about the value of the 6th amendment, about what it means to treat um what it means to treat people who are in the crosshairs of the criminal justice system, but rather than having that conversation, uh oh, it's just this consumerism nation model. Well, there's a lot of noise out here, so we're going to react in this sort of way. Higher education as well, unfortunately, has been commodified and other sorts of ways that has reduced or or impeded hampered uh these schools commitments to uh free and robust and open dialogue. So, to the degree that academic freedom doesn't sit squarely at the center of the academic mission, uh any school was gonna be in trouble and I really hope that uh um that we weather this current mhm political moment where uh 19 year olds without degrees or running universities and get back to a a system where uh faculty where adults make decisions in the best interests of the university and the best interests of the students even to the degree, though some of those decisions may be uh unpopular and that uh is going to require a certain courage. Um and hopefully in time and I'm confident that in time, um administrators are going to begin to push back on these current trends. Harvard's been around for a long time, it's been around for a long time for a reason. And one of the reasons is that it understands itself not to be static. So I have every uh view that Harvard is going to adapt and Get itself back on course and be around another 400 years at least. That's my </s><s>[INST] so I mean what does it kind of boils down to is just having difficult conversations, difficult debates uh when you mentioned sort of 19 year olds and it's funny I've seen this unit at M. I. T. It's not that they shouldn't have a voice should they do seem to I guess you have to experience it and just observe it. They have a strangely disproportionate power. It's very interesting to to basically, I mean you say yes there's great faculty and so on. But you know uh it's not even just that the faculty is smart or wise or whatever it is that they're just silenced. So the terminology that you mentioned is weaponized as a sort of safe spaces or that certain conversations made people feel unsafe. What do you think about this kind of idea? You know? Is there some things that are unsafe to talk about in the university setting? Is their lines to be drawn somewhere? And uh just like you said on the flip side with a slippery slope, is it too easy for the lines to be drawn everywhere? [/INST] Yeah that's a great question. So this idea of unsafe space at least the vocabulary derives from some research academic research about feeling psychologically uh unsafe. And so the notion here is that there is uh there are forms of uh psychological disquiet that impedes people from experiencing the educational environment to the greatest degree possible. And that's the argument. I and and assuming for a moment that people do have these feelings of of of disquiet uh elite universities like M. I. T. And like Harvard, that's probably the safest space people are gonna be in for their their their lives. Because when they get out into the quote unquote real world, uh they won't have the sorts of nets that these schools provide safety nets that these schools provide. So to the extent that research is descriptive of a psychological feeling, I think that the duty of the universities are to challenge people. It seems to me that it's a shame to go to a place like Harvard, a place like M. I. T. Yell any of these great institutions and come out the same person that you were when you went in uh that seems to be a horrible waste of four years and and and money and and and resources rather, we ought to challenge students that they grow um challenge some of the uh their most deeply held assumptions. Um they may continue to hold them, but the point of an education is to rigorously interrogate um these fundamental assumptions that have guided you uh thus far and to do it uh fairly and civilly. So the extent that there are lines that should be drawn. There's a long tradition in the University of civil discourse. So you should uh drawing lines somewhere between civil discourse and uncivil discourse, the purpose of a university is to talk difficult conversations, tough issues uh talk directly and frankly, but do it civilly, and you know, so to um you know, yelling cuss at somebody uh and that sort of thing. Well, you know, do that on your own space, but observe the norms of civil discourse at at the university. Uh So look, I think that the presumption ought to be that uh the most difficult topics are appropriate to talk about at a universe, that that ought to be the presumption. Now, you know, I should uh um m I t for example, give it's prom imprimatur to someone who is espousing uh the flat Earth theory, you know, the Earth is flat, right? So they're if if certain ideas uh are are so uh contrary to the scientific uh and and and cultural thinking of the of the moment. Yeah, there's space there to draw a line and say, yeah, we're not going to give you this platform to tell our students that the Earth is flat. Uh but, you know, a topic that's controversial but contested story, that's what universities are for if you don't like the idea, present better ideas and articulate them, </s> |
<s>[INST] smooth pickup line by the way, if I've ever heard one that's good, you want to get out of here, [/INST] so start dating and she was the girl that screwed my brains out man and I felt head over hills, we got married six months later, six months, that's what it does and I had, I was, she didn't know it was a crook. She had no idea. Now. She knew I was very bright, you know, I would do a lot of theater stuff like that. I got a job at, I was in Hazzard, there was no jobs to be had. So I got a job in Lexington because we were going to be moving to the UK, got a job in Lexington at Lexmark testing, printer boards, um, circuit boards. So I would leave on a thursday night work 3 18 hour shifts at Lexmark, come back home on monday. Um, got married, faked a car accident to get that. Then the other, the rest of the money that I needed to get married. And the faking on that man. I had bought a A Chevy spectrum at a car auction gave like $500 for it. My aunt had previously defrauded usa insurance on a car accident and she was telling me all about it. She's like, look, go down to this chiropractor, make sure you get the insurance where they'll pay for a rental car, they'll pay lost wages. I was like, they pay lost wages. She's like, yeah, they pay lost wages. I was like, she's like, by the way you work for me and I was like, I work for you. So </s><s>[INST] you don't get to define the with the wage. And you can also define how long you were unable to work. [/INST] Exactly. Exactly. And the chiropractor will sign off on any damn thing. Alright, so my cousin Ronnie, he figures out that I'm going, he finds out I'm going to fake this car accident. So he comes to me, he's like, hey man, can I get in on that? I was like, yeah man, you get on that. So this kid, he's five days younger than I am, this kid. He goes to the dentist the day that we're faking, it has a tooth pulled, tells the dentist not to numb it, not to stitch it. Just pull it. So he shows </s><s>[INST] up, [/INST] he shows up the day that we're driving out to fake the accident. He's got blood all over his shirt. He's still bleeding out of the mouth and everything else. I'm like, are you okay? And he's like, yeah man, it's gonna be good. It's gonna be good. I'm like, ok. So my mom, by this point, I'm living with my grandparents. My mom is up in the head of a hollow. So we're like, well just do it up there. We'll go act like we're visiting my mom on the way back out, ran over a mountain. Okay? So we go visit and everything come back out that night, run over the side of the hill. Me and Ronnie walked back up. Of course it totals the car, walk back to my mom's acting like we've wrecked. She knows what time it is and everything else and follow the claims. So that gets the money to get married and me and my wife moved from hazard to Lexington and I'm the kid that my crime usually if I was a single guy wouldn't break the law and I would be all right, you know, but females involved Oh yeah, Oh yeah, I got to spend the money. You gotta show them gifts. Everything else was never enough to show love in some sort of healthy way. Always had to go overboard and typically it was buying or stealing some sort of expensive. </s><s>[INST] So that, that was, that was the thing that was the way you show love is by buying expensive gifts [/INST] or something overboard back then with with Susan initially it was don't worry about working. I got it. You just worry about going to school. She was a music major. I was like, you just worry about going to school, so don't worry about cooking and cleaning. I got it. I got it. So not only was I this this guy that was going overboard, but it's kind of a control freak too. Right? So now I got it. I got it. I got it. So here I am, you know, 60 hour week job, 18 hour class load, cooking, cleaning, something had to give, I quit the job. I couldn't do it, quit the job and start back in fraud and trying to hide that from her at the same time. So it was initially telemarketing fraud started, I was working at the first job I had was a telemarketer at a cemetery selling gravesites and then that ended, went over to work for the Shriners Circus, Shriners Hospital and there was a third party company that was doing all the telemarketing, made really good money doing that. That job ended and then they pivoted over to um working with Kiwanis clubs selling food baskets to the food banks and everything. So I stole the phone list and started at my own Kiwanis club and would do the telemarketing, go out twice a week and pick up checks. Well what happened was is I'm going out picking up Jax, go knock on the door, turns out one of the persons that I had called was a law enforcement officer, so he was like who are you? I'm like I'm with the Kiwanis club and he's like no you're not. </s><s>[INST] So [/INST] got arrested, spent three months in the county jail for theft by deception. Got out and we had to move from from Lexington back to Hazzard and lived with Susan's parents. They had gotten a desktop computer hp and I started surfing around online, found Ebay and didn't really know how to make money on Ebay about the same time, I'm committing low level frauds online, I don't really talk about that in the past the first time I've really talked about that but I would pay for it with bad checks. So </s> |
<s>[INST] And that detector then converts that into an electrical signal. And the model signal is correlated with some kind of weight. [/INST] So the synaptic weight will tell you how many flux on to pop into the loop. It's an analog number. We're doing analog computation. Now, </s><s>[INST] can you just linger on what the heck is the flux on? I was supposed to know. This is this is this is a funny, it's like the big bang. Is this a funny word for something deeply technical. [/INST] No, let's let's try to avoid using the word flux because it's not actually necessary when a when a when a photo fun to </s><s>[INST] say though. So uh so it's very necessary, I would say [/INST] when a when a photon hits that superconducting single photon detector current is added to a superconducting loop. And the the amount of current that you add is an analogue value, can have eight bit equivalent resolution, something like that. 10 bits maybe. </s><s>[INST] That's amazing. By the way, this is starting to make a lot more sense [/INST] when you're using superconductors for this. The energy of that circulating current is is less than the energy of that photons. So your energy budget is not destroyed by doing this analog computation. So now in the language of of a neuroscientist, you would say that's your post synaptic signal. You have this current being stored in a loop, You can decide what you want to do with it. Most likely you're going to have a decay exponentially. So every single synapse is going to have some given time constant. And that's determined by set by putting some resistor in that, in that superconducting loop. So a synapse uh synapse event occurs when a photon strikes a detector adds current to that loop. It decays over time. That's the post synaptic signal, then you can process that in a dendritic tree bryce primavera and I have a paper that we've submitted about that um For the more neuroscience oriented people there's a lot of dendritic processing, a lot of plasticity mechanisms you can implement with essentially exactly the same circuits. You have this one simple building blocks circuit that you can use for a synapse. For a dendrite, for the neuron cell body for all the plasticity functions. It's all based on the same building block, just tweaking a couple </s><s>[INST] parameters. This basic building block has both an optical and electrical component and then you can just build arbitrary large systems with that [/INST] close you're not at fault for thinking that that's what I meant. What I what I should say is that if you wanted to be a synapse you tack a detector, a superconducting detector onto the front of it. And if you wanted to be anything else there's no optical component. Got it at the front </s><s>[INST] optics in the front, electrical stuff in the back [/INST] electrical yeah in the processing and in the the output signal that it sends to the next stage of processing further, </s><s>[INST] the dendritic trees is electrical. [/INST] It's all electrical. It's all electrical in the super cannon domain. For anybody who's up on their superconducting circuits. It's just based on a D. C. Squid. The most ubiquitous. It's which is a circuit composed of two Josephson junctions. So it's it's a very bread and butter kind of thing. And then the only place where you go beyond that is the neuron cell body itself. It's receiving all these electrical inputs from the synapses or dendrites or however you've structured that particular unique neuron. And when it reaches its threshold which occurs by driving a Josephson junction above its critical current, it produces a pulse of current which starts an amplification sequence voltage amplification that produces light out of a transmitter. So one of one of our colleagues, adam Akane and Sonia buckley as well, did a lot of work on the the light sources and the amplifiers that drive the current and produce sufficient voltage to drive current through that Now semiconducting part. So that light source is the semiconducting parts of a neuron and that so the neuron has reached threshold. It produces a pulse of light that light then fans out across the network of wave guides to reach all the downstream synaptic terminals that perform this process themselves. So it's probably worth explaining what a network of wave guides is because a lot of listeners are going to know that look up the papers by Jeff Child's on this one. But basically light can be guided in a simple, basically wire of usually an insulating material. So silicon silicon nitride. Different different kinds of glass, just like in a fiber optic, it's glass. Silicon dioxide that makes it a little bit big. We want to bring these down. So we use different materials like silicon nitride but basically just imagine a rectangle of some material that just goes and branches um forms different different branch points that target different subregions of the network, you can transition between layers of these. So now we're talking about building in the third dimension, which is absolutely crucial. So that's what wave guides are. </s><s>[INST] Yeah, that's great. What? Why the third dimension is crucial. [/INST] Okay, so yes, you were talking about? What are some of the technical limitations? one of the things that I believe we have to grapple with is that our brains are miraculously compact for the number of neurons that are in our brain. It sure does fit in a small volume as you would have to if we're going to be biological organisms that are resource limited and things like that. Any kind of hardware neuron is almost certainly going to be much bigger than that. If it is of comparable complexity. Even whether it's based on silicon transistors. Okay, a transistor, seven nm, that doesn't mean semiconductor based neuron is seven nanometers. They're big, they require many transistors. Different other things like capacitors and things that store charge. They end up being on the order of 100 microns by 100 microns and it's difficult to get them down any smaller than that. The same is true for superconducting neurons and the same is true if we're trying to use light for communication, even if you're using electrons for communication, you you have these wires where Okay, the the size of an electron might be angstrom is but the size of a wire is not angstrom. And if you try and make it narrower, the resistance just goes up. So you don't actually win to communicate over long distances. You need your wires to be microns wide. And it's the same thing for wave guides. Wave guides are essentially limited by the wavelength of light and that's going to be about a microphone. So whereas compare that to an ax on the analogous component in the brain, which is 10 nm in diameter, something like that. The they're bigger when they need to communicate over long distances. But grappling with the size of these structures is inevitable and crucial. And so in order to make systems of comparable scale to the human brain by scale here. I mean number of interconnected neurons you absolutely have to be using the third spatial dimension. And that means on the way for you need multiple layers of both active and passive components active. I mean superconducting electronic circuits that are performing computations and passive. I mean these wave guides that are routing the optical signals to different places. You have to be able to stack those. If you can get to something like 10 planes of each of those or maybe not even 10, maybe 56, something like that. Then you're in business. Now you can get millions of neurons on a wafer but that's not that's not anywhere close to the brain scale. In order to get to the scale of the human brain, you're going to have to also use that their dimension in the sense that entire wafers need to be stacked on top of each other with fiber optic communication between them. And we need to be able to fill a space the size of this table with stacked wafers. And that's when you can get to some 10 billion neurons like the human brain. And I don't think that's specific to the Opto electronic approach that we're taking. I think that applies to any hardware where you're trying to reach commensurate scale and complexity of the human brain. </s> |
<s>[INST] are in a negative frame of mind. So the natural ability of a Chad bot to be positive is just going to give you a higher success rate. [/INST] Yeah, yeah. And they're not gonna get mad and argue with you. You know, you say to a Chad bot, you know, your price is too high. Chad bot is designed to come back with a smiley face. Yeah. You say to a person, your price is too high. They go, how dare you? I'm trying to make a living. </s><s>[INST] You know, they gotta go off the deepest. Unfortunately or fortunately, I think the way Chad bots are going now, they will come back negative because they're becoming more and more humanlike. That's the whole point. To be able to pass the touring tests, you have to be negative, you have to be an asshole, you have to have boundaries, you have to be insecure, [/INST] you have to have some uncertainty. Well, as them pursuing being having boundaries and being negative. Like I can, again, you throw a proposal to me, Before I say no, I'm going to say, look, I'm sorry, that just doesn't work for me. I'm going to set up a real clear boundary without being negative. A lot of people really struggle with setting boundaries without being negative, without name calling, without indignation, </s><s>[INST] without getting upset. When you show that you're not getting upset, I'm not just seeing that. I'm seeing a flawed human that has underneath it a temper, underneath it the ability to get upset, but chooses not to get upset. And the chatbot has to demonstrate that. So it's not just going to be cold and be this corporate, blank, empty, vapid creature that just says, oh, thank you. Thank you for saying that. No. The chatbot has to be able to be mean [/INST] and choose not to be. Interesting. I don't know. I'd be willing to see that </s><s>[INST] play out and see how it plays out. Maybe not, but I guess what I'm saying is to be a good negotiator, you have to have the capacity to be a bad person and choose not to. Really? I think so. Really? See, I think you're just going to have the capacity to set a boundary and stick to it. Interesting. Because I think it's hard for me to trust a person who's not aware of their own demons, because if you say you don't have any demons, if you don't have any flaws, I can't trust you. Yeah. Well, it's first of all, it's a lie, right? So somebody's life, right? This is gives back to life Yes, so you have to have a self-awareness about that But he'd be able to control and demonstrate the able to control it. I mean this is humans. I just think humans Intelligent effective humans are able to do this. Well and chat bots are not Yet and they're moving that direction. So it makes me think about what is actually required for effective negotiation That's what AI systems do is they make you ask yourself What is it that makes humans special any discipline? What is it that makes humans special chess and go games? Which AI systems are able to beat humans at now. What is it that makes them effective at? Negotiation, what does it make them effective at? something that's extremely difficult which is Navigating physical spaces. So doing things that we take for granted like making yourself a cup of coffee is exceptionally difficult problem For robots. Yeah, because of all the complexities involved in navigating physical reality We we have so much common sense reasoning built in just about how gravity works about how Objects move the What kind of objects there are in the world It's like it's it's really difficult to describe because it all seems so damn trivial But it's not trivial right because a lot of that we just learn as babies We keep running into things and we'll learn about that So AI systems help us understand. What is it that makes humans release? What is the wisdom we have in our heads and And negotiation to me is super interesting because negotiation is not it's about business. It's about Geopolitics it's about running government Is basically negotiating? How do we have the different policies different? Bills and programs and so on how do we allocate money? How do we reallocate resources all that kind of stuff that seems like AI in the future could be better at that right, but maybe not maybe you have to be a messy weird insecure Uncertain human and debate each other and yell at each other on Twitter Maybe you have to have the red and the blue teams that yell at each other In in the process of figuring out what is true. Maybe AI systems will not be able to do that [/INST] I figure out the full mess of human of human civilization. Yeah, interesting Well, I mean the two thoughts that I had along the way was I mean is anytime you're talking about systems or scaling You know, yeah, you're talking my belief is chatbot systems things that don't require Decision-making just following the instructions at least 80% of what's going on now The remaining percentage whatever it is. Does it does it require the human interaction and what's required? Like I'm not I'm not I'm not like I am NOT pro conflict And I also know that there's a case to be made in the creative world. It's some of the best thinking came out of conflict I'm reading interviews of Bono you too you know their admiration for some of the Beatles best music came when they were fighting with each other and The song one octune bay, which is I believe from the album octune baby Those guys were fighting. I mean they were on a verge of breaking up and their appreciation that conflict could create something beautiful And then when I was in a crisis negotiation unit, you know my last seven years in the FBI There was a guy that um named Vince brain dude brain brain negotiator And he and I used to argue all the time And then when we had a change in the guy who was in charge the guy was in charge took me off to the side He's like, you know, I can't take you and Vince fight not and I said, well, I got news for you I think we come up with much better stuff as a result of our battles And he said, you know Vince said the same thing to me and I like so if we don't have a problem fighting Why do you have a problem, but you know some there's there is something there that sometimes the most difficult? insights You rack your brains as to why someone is so Dug in on something that so you thinks is so wrong. Yeah, maybe there's something to it </s> |
<s>[INST] detect those. So how early is this work in a maybe in the history of science. So I mean like uh so if you were to if if you and I time travel 1000 years into the future uh organized seemed to be maybe I'm romanticizing the notion but you're building not a brain but something that has properties of a brain. So it feels like you might be getting close to in the building process to build to understand. So how um how far are we in this understanding process of development, [/INST] 1000 years from now? It's a long time from now. So if this planet is still gonna be here 1000 years from now. So </s><s>[INST] I mean if you know like they write a book, obviously there'll be a chapter about the science [/INST] fiction book today </s><s>[INST] today, I guess, where we really understood very little about the brain a century ago. Well, I was a big fan in high school, reading Freud and so on, and still am of psychiatry, I would say we still understand very little about the functional aspect of just but how in the history of understanding the biology of the brain, the development, How far are we along? [/INST] It's a very good question. And so this is just of course, my opinion. Um I think that we did not have technology um even 10 years ago or 20, certainly not 20 years ago to even think about experimentally investigating the development of the human brain. So, we've done a lot of work in science to study the brain or many other organisms. Now we have some technologies which spell out that allow us to actually look at the real thing and look at the brain at the human brain. So, what are these technologies? There has been a huge progress in stem cell biology the moment someone figured out how to turn a skin cell into an embryonic stem cell, basically, and that's how that embryonic stem cells could begin a process of development again, to, for example, make a brain, there was a huge and you know, advance and in fact, there was a Nobel prize for that, that started the field really of using stem cells to build organs. Now, we can build on all the knowledge of development. We build over the many, many, many years to say, how do we make these stem cells now, make more and more complex aspects of development of the human brain? Uh, so this field is younger, the field of brain organize, but it's moving faster and it's moving fast in a very serious way that is rooted in labs with the right ethical framework and um and really building on, you know, solid science for realities and what is not and um, but it will go faster and it will be more and more powerful. We also have technology that allows us to basically study the properties of single cells across many, many millions of single cells, which we didn't have perhaps five years ago. So now with that even an organ oy that has millions of cells can be profiled in a way, looked at very, very high resolution, the single cell level to really understand what is going on and you could do it in multiple stages of development and you can build your hypothesis and so on and so forth. So it's not going to be 1000 years, it's gonna be a shorter amount of time. And I see this as a sort of a an exponential growth of this field enabled by these technologies that we didn't have before. And so we're going to see something transformative that we didn't see it all in the prayer 1000 years. </s><s>[INST] So I apologize for the crazy sci fi questions, but uh, the development and process is fascinating to watch and study, but how far are we away from and maybe how difficult is it to build? Not just an organ oid but a human brain. Okay. From a stem cell. Yeah. [/INST] First of all, that's not the goal for the majority of the serial scientists that work on this, because you don't have to build the whole human brain to make this model useful for understanding how the brain develops or understanding disease. You don't have to build the whole thing. </s><s>[INST] So let me, let me just comment on that is fascinating. It shows to me the difference between you and I as you're actually trying to understand the beauty of the human brain and to use it to really help thousands or millions of people with disease and so on. Right. Uh from an artificial intelligence perspective, we're trying to build systems that we can put in robots and try to create systems that have echoes of the intelligence about reasoning about the world, navigating the world. It's different objectives. I think [/INST] that's very much science fiction, science </s><s>[INST] fiction, but we operate in science fiction a little bit. But so so on that point of building a brain, even though that is not the focus or interest perhaps of the community, how difficult is it? Is it truly science fiction at this point? [/INST] I think the field will progress like I said, and that um this system will be more and more complex in a way, right? But there are properties um, that emerged from the human brain that have to do with the mind that may have to do with consciousness, may have to do with uh, intelligence or whatever, that we don't really don't don't understand even how they can emerge from an actual real brain and therefore we cannot measure or study in an organized. So, um, so I think that this field, many, many years from now may lead to the building of better neural circuits of uh, you know, that that really are built out of understanding of how this process really works. And it's hard to predict how complex this really will be. I really don't think we're so far from. It makes me laugh, really? It's really that far from building the human, the human brain, but you're going to be building something that is, um, you know, always a bad version of it, but that may have really powerful properties and might be able to, you know, respond to stimuli or or be used in in certain context. And this is why I really think that there is no other way to do this science, but within the right ethical framework. Because where you're going with this is also, you know, we can talk about science fiction and write that book and we could today. But this work happens in a specific ethical framework that we don't decide just a scientist, but also as a society. So, </s> |
<s>[INST] if if I [/INST] can do that again, maybe I never become, you know, high level black belt competitor, but you know, I had four of the things I was doing my life. I also have a family, I have this, I have that, you know what that person was able to accomplish in the martial arts relative to what they were able to put in this phenomenal, you know, other times someone could be a very successful black belt in my mind, be a bump because they could have been a lot more and you know, they could have done more, they could have focused more and there's no shame in deciding that you don't want to do that, but whatever it is that you're, you're invested in, I remember the uh take it uneasy podcast and that I love because you know, I'll just chill out like resting. It's like vacation who wants to go on vacation, I'll go on vacation for a day or two. You want to spend three weeks on vacation. Like I kill myself like get me out of here. Like this is part of this is I'm a waste of life. I'm not doing anything useful. Your </s><s>[INST] technical application right now. [/INST] Right, Well, this is fun though, but it's like a one day vacation. Exactly. But you know, if you had, I'm sure you're thinking about jumping off of the building right now. But if you had, if you had to talk to me for, you know, like three days, I'm sure you do. You probably shut me off the building. I don't blame you, I'll be done, but five hours in. But yeah, but you know, it's like you want to be pushing towards something because otherwise, what's the purpose of being here? You know, it's not just a college, it's doing something useful building growing as a person, helping others do the same if that's within your power at any given time. But I think that's kind of the neat thing about martial arts is it can be many, many different things to many different people, you know, I finally, for instance, was able to get a college degree this this year. That, which I mean, it's not a big deal for most people, but for me it was a big deal because I </s><s>[INST] was going back and finish. [/INST] Yeah. And I never envisioned ever going back and </s><s>[INST] that's a hard step to go back and finish. That's uh it weighs heavy on you if you don't, it's interesting. [/INST] I was just, I was more proud of that than most things I've ever done if I'm honest, you know, and it was neat and I really enjoyed it and it was the process of doing it, but you know, are my academic credentials impressive, like not in the least, but for me it's like it was a big deal for me personally to take that step and to go back and do that and I was proud of the direction and because it would have been easy, like do I need to do it, like no, you know, business, I'll do, okay, I'll try, I'll keep fighting, but I was happy to take the time in between fights when I was when I was wasn't booked for an opponent to do something productive rather than just, I'll just hang out, you know, like I can still train every single day, but I can also train and go to school. People go to the olympics while going to school. I can, I can do martial arts and and go to school. </s><s>[INST] Oh one thing I gotta ask is uh you know, a bunch of women listen to this podcast if they haven't done the jets who think it be kind of intimidating to step on the mat with a bunch of browse uh that like enjoy somehow killing each other. Like how do you succeed in that environment to where you can learn this art, learned how to beat all those people up. [/INST] Um Oh gosh, </s><s>[INST] is there any advice? I mean another way to ask that is like if, if uh any women listening to this are interested in starting digits. Like is there advice for that journey? [/INST] Honestly, I think it's just walking in the door and starting sometimes. I don't know how to respond because I'm not, I don't view myself as typically anxious particularly um uh interactions with other people or new people. Shai Shai is not a word that has been used for me, but if you ask my family and um they joke because our son talks a lot, he's advanced family and they're always like, oh well we know where he gets that from, like because he just doesn't stop talking. He narrates everything he does. Um and so they always tease because that's like I'm known for kind of talking a lot. Um but so I haven't been, typically, I'm not, I don't consider myself a shy person, so for me going into um a new room, a new group of people is, you know, there's always that you don't really know who they are, how they're going to treat you, but but I, I don't have a lot of anxiety with that, so I don't know if that's something that's going to put something up, I don't really know how to address that particular feeling. Um but in terms of all of the rooms I've been and I have popped into jesse James before, I knew Ryan in floor, like I traveled for my job, so in Germany in florida and California in places where I don't know anyone, they don't know me and I have never once had anyone be anything other than, than kind and solicitous and helpful um and long before, when I was a white belt and a bluebell and didn't know anything and I didn't know anyone and um I just think that is a community of people that it's so cool that no matter where you go in the world, um I walked into a gym in Prague one time where only two people spoke english and it was just, yeah, it's weird, </s> |
<s>[INST] totally agree with you. Unfortunately his job was very tough and all those kinds of things, but um, I see arrogance and arrogance prevented him from being honest in that way previously and I think arrogance will prevent him from being honest in that way. Now we need leaders and I think young people are seeing that that kind of talking down to people from a position of power. I'm, I hope is the way of the past. People really like authenticity and they, they like leaders there like a man and a woman of the people and I think, [/INST] I mean he still has a chance to do that. I think, I </s><s>[INST] mean, I [/INST] don't think, you know if I doubt he's listening, but if he is like, hey, I, I think, you know, I don't think he's irredeemable by any means. I think, you know, I don't, I don't have an opinion on whether there was arrogance or there or not. Um, just know that I think like coming clean on the, you know, it's understandable to have sucked up during this pandemic. Like I don't expect any government to handle it well because it was so difficult. Like so many moving pieces, so much like lack of information and so on. Um, but the step to rebuilding trust is to go, okay, look, we're doing a scrutiny of where we went wrong and I, and for my part, I did this wrong in this part and that would be huge. </s><s>[INST] All of us can do that. I mean, I was struggling for a while whether I want to talk to him or not, I talked to his boss, Francis, Collins. Um, another person that's got screwed up in terms of trust. Um, I lost a little bit of my respect to. There seems to have been a kind of dishonesty in the, in the back rooms, in that they didn't trust people to be intelligent. Like we need to tell them what's good for them. We know what's good for them. That kind of idea [/INST] to be fair. The thing that's what's it called, the phrase today. Not picking. Social media does that. So you've got like, nit picking, Not picking is where the craziest, stupidest. You know, if you have a group of people, let's call, you know, let's say people who are vaccine. I don't like the term anti vaccine. People who are vaccine hesitant vaccine speculative, you know what social media did or the media or anyone, you know, their opponents would do is pick the craziest example. So the ones who are like, you know, I think I need to inject myself with motor oil up my ax or something, </s><s>[INST] you know, [/INST] select the craziest ones and then have that beamed to, you know, so from like someone like Fauci or Francis's perspective, that's what they get because they're getting the same social media stuff as us. They get the same media reports. I mean they might get some more information, but they they too are going to get these the nuts portrayed to them. So they probably have a misrepresentation of what the actual public's intelligence is. That </s><s>[INST] that just Yes. And that just means they're not social media savvy. So one of the skills of being on social media is to be able to filter that in your mind. Like to understand, to put in proper content, [/INST] realize that what you are seeing social media is not anywhere near an accurate representation of humanity. </s><s>[INST] Not picking and there's nothing wrong with putting more oil up your assets just as I do this every weekend. Okay. [/INST] That analogy come from in my mind. Like what </s><s>[INST] I don't know. I think we need to, there's some Freudian thing would need to deeply investigate with a therapist. Okay. What about ai are you worried about A. G. I. Super intelligence systems or paperclip maximizing type of type of situation? [/INST] Yes, I'm definitely worried about it. But I feel kind of bipolar and some days I wake up and I'm like </s><s>[INST] you're excited about the future. [/INST] Well exactly. I'm like wow we can unlock the mysteries of the universe, you know, escape the game. Um And this this you know, because I spend all my time thinking about these monarchy problems that you know what what is the solution to them? You know in some ways you need this like omni benevolent, omniscient omni wise coordination mechanism that can like make us all not do the monarchy thing or like provide the infrastructure would redesign the system so that it's not vulnerable to this monarchy process. Um And in some ways, you know that's that's the strongest argument to me for like the race to build A G. I. Is that maybe, you know, we can't survive without it. But the flip side to that is the unfortunately now that there's multiple actors trying to build A. I. A. G. I. You know this is this is fine 10 years ago when it was just Deepmind. But then other companies started up and now it created a race dynamic. Now it's like that's the whole thing is that the same? It's got the same problem. It's like whichever company is the one that like optimizes for speed at the cost of safety will get the competitive advantage. And so will be the more likely the ones to build the A. G. I. You know, and and that's the same cycle that you're in. And there's no clear solution to that because you can't just go like um slapping the if you go and try and like stop all the different companies, then it will, you know, the good ones will stop because they're the ones, you know, within, you know, within the West's reach. But then that leaves all the other ones to continue. And then they're even more likely. So it's like, it's a very difficult problem with no clean solution. Um and, you know, at the same time, you know, I I know the at least some of the folks at deepmind and they're incredible and they're thinking about this, they're very aware of this problem. And they're like, you know, I think some of the smartest people on earth </s> |
<s>[INST] first of all I wonder if some of that is because we develop our intelligence and our emotional intelligence by communicating with other humans. And so one other humans disappear from the picture was still kind of a virtual human. The [/INST] code still runs, </s><s>[INST] The code still runs and but you also kind of you're still there's like virtual humans, you don't have to think of it that way. But there's a kind of when you like chocolate. Yeah like you're you're kind of chuckling to a virtual human. I mean it's possible that the code is has to have another human uh there because if you just grew up alone, I wonder if the motion will still be there in this visual form. So yeah I wonder but anyway the what can you tell from the human face about what's going on inside. So that's the problem that affect your first tackled which is using computer vision using machine learning to try to detect stuff about the human face as many things as possible and convert them into a prediction of categories of emotion anger, happiness, all that kind of stuff, how hard is that problem? [/INST] Extremely hard. It's very, very hard because there is no 1-1 mapping between um a facial expression and your internal state. There just isn't there's this over simplification of the problem where it's something like if you are smiling then you're happy. If you do a brow furrow then you're angry. If you do an eyebrow raise then you're surprised and just think about it for a moment. You could be smiling for a whole host of reasons. You could also be happy and not be smiling, right? Um you could follow your eyebrows because you're angry or you're confused about something or you're constipated. Um so I think this over simplistic approach to inferring emotion from a facial expression is really dangerous. Um the solution is to incorporate as many contextual signals as you can, right? So if for example, I'm driving a car and you can see me like nodding my head and my eyes are closed and the blinking rate is changing, I'm probably falling asleep at the wheel right? It doesn't because you know the context you understand what the person is doing. So I think or add additional channels like voice or gestures or even physiological sensors. Um but I think it's very dangerous to just take this over simplistic approach of Yeah, smile equals happy and </s><s>[INST] if you're able to, in a high resolution way specify the context, There's certain things that are going to be somewhat reliable signals of something like drowsiness or happiness or stuff like that. I mean when when people are watching netflix, that problem, that's a really compelling idea that you can kind of at least an aggregate highlight, like which part was boring, which part was exciting, how hard was that problem? [/INST] Um That was on the scale of like difficulty, I think that's some of one of the easier problems to solve because it's um um relatively constrained environment. You have somebody sitting in front of. Initially we started with like a device in front of you, like laptop and then we graduated to doing this on a mobile phone, which is a lot harder just because of, you know, from a computer vision perspective, the profile view of the face can be a lot more challenging. We have to figure out lighting conditions because usually people are watching content literally in their bedrooms at night, lights are dimmed um </s><s>[INST] Yeah, I mean if you're standing it's probably gonna be the looking up [/INST] the nostril view. </s><s>[INST] Yeah, and nobody looks good at. I've seen data sets from that perspective, it's like this is not a good look for anyone or if you're laying in bed at night, what is it? Side view or something. It's and half your face is like on a pillow. Actually, I would love to know have data about uh like how people watch stuff in bed at night, like a prop, there isn't a pillow. The like I'm sure there's a lot of interesting dynamics [/INST] from health and well being perspective, right? Like it's like you're </s><s>[INST] learning perspective but yes but also yeah, yeah once you have that data you can start making all kinds of influence about health and stuff like [/INST] that interesting. </s><s>[INST] Yeah there's an interesting thing when I was a google that we were uh it's it's called active authentication where you you want to be able to unlock your phone without using a password so it would it would face but also other stuff like the way you take the phone out of the pocket. So that kind of data to use the multi moto with machine learning to be able to identify that it's you or likely to be you likely not to be you that allows you to not always have to enter the password. That was the idea. But the funny thing about that is I just want to tell a small anecdote is because it was all male engineers. Uh except so there's, my boss is, our boss was still one of my favorite humans was was a woman Regina dugan. Um Oh [/INST] my God, I love her, </s><s>[INST] she's awesome. Um So uh but anyway there was another, there's one female engineer, brilliant female engineer on the team and she was the one that actually I like the fact that um women often don't have pockets. [/INST] It </s> |
<s>[INST] Right. That's a fascinating question to ask of a book. [/INST] You can do that. Right? So some paragraphs could be reordered, some sentences can be reordered, you could say he is tall and smart and X. Right? And it doesn't matter the order of Fallen Smart. But if you say the tall man is wearing a red shirt, what colors? You know, like you can create dependencies. Right. Right. And so GPS on the other hand, run simple programs on pixels, but you're given a million of them and the first order the screen you're looking at. It doesn't care which order you do it in. So I call that given parallelism simple narratives around the large numbers of things where you can just say it's parallel because you told me it was </s><s>[INST] so found parallelism where the narrative is sequential, but you discover like little pockets of parallelism versus [/INST] turns out large pockets of parallelism </s><s>[INST] large. So how hard is it to this? [/INST] How hard is it that's just transistor count. So once you crack the problem, you say, Here's how you fetch 10 instructions at a time, here's how you calculated the dependencies between them. Here's how you describe the dependencies, here's, you know, these are pieces. Right? </s><s>[INST] So once you describe the dependencies, then it's just a graph sort of it's an algorithm that finds what is that? I'm sure there's a graph, their theoretical answer here, that's solvable in general programs modern programs that human beings. Right? How much Found parallelism? Is there? What does 10 next mean [/INST] you executed in order </s><s>[INST] versus [/INST] you would get what's called cycles per instruction and it would be about, you know, three instructions, three cycles per instruction because of the latency of the operations and stuff. And in a modern computer, excuse that, like point to .25 cycles per instruction. So it's about We today find 10 x. And then there's two things. One is the found parallelism and the narrative, right? And the other is to predictability of the narrative. Right? So certain operations, they do a bunch of calculations and if greater than one do this else do that. That that decision is predicted in modern computers to high 90% accuracy. So branches happen a lot. So imagine you have, you have a decision to make every six instructions which is about the average. Right? But you want to fetch 500 instructions. Figure out the graph and execute tomorrow parallel that means You have, let's say if you affect 600 instructions and it's every six You have to fetch, you have to predict 99 out of 100 branches correctly for that window to be effective. </s><s>[INST] Okay, so parallelism, you can't paralyze branches. You can, [/INST] you </s><s>[INST] can what is predicted branch mean [/INST] what's predicted? Imagine you do a computation over and over. You're in a loop. So while and is greater than one do and you go through that loop a million times. So every time you look at the branch you say it's probably still greater than one. </s><s>[INST] You're saying you could do that accurately. [/INST] Very accurately monitor </s><s>[INST] my mind is blown. How the heck do you do that? Wait a minute. [/INST] Well you want to know this is really sad. 20 years ago you simply recorded which way the branch went last time and predicted the same thing. Right, okay, </s><s>[INST] what's the accuracy of that, [/INST] 85%. So then somebody said, hey let's keep a couple of bits and have a little counter. So it predicts one way we count up and then pins so save a three bit counter. So you count up and then count down and if it's, you know, you can use the top, it is the sign bit. So you have a signed two bit number. So if it's greater than one, you predict taken and less than one you predict not taken right or lessons or whatever the thing is And that got us to 92%. Okay. I know it is better. This branch depends on how you got there. So if you came down the code one way, you're talking about bob and jane, right? And then said it's just bob like jane, it went one way. But if you're talking about bob and Jill does bob like changes you go a different way. Right. So that's called history. So you take the history and encounter. Mhm. That's cool. But that's not how anything works today. They use something that looks a little like a neural network. So modern, you take all the execution flows and then you do basically deep pattern recognition of how the program is executing and you do that multiple different ways and you have something that chooses what the best result is. There's a little supercomputer inside the computer calculates which way branches go. So the effective window that it's worth finding grass and gets bigger. </s><s>[INST] Why was that gonna make me sad because that's amazing. [/INST] It's amazingly complicated. Well, here's the here's the funny thing so to get to 85% took 1000 bits to get to 99% takes tens of megabits. So this is one of those To get the result, you know, to get from a window of say 50 instructions to 500. It took three orders of magnitude or four orders of magnitude more bits. </s><s>[INST] Now, if you get the prediction of a branch wrong, what happens then flush the pipe to flush the pipes is just the performance cost, [/INST] but it gets even better. So we're starting to look at stuff that says so executed down this path and then you have two ways to go. But far far away there's something that doesn't matter which path you went. So you miss, you took the wrong path, you executed a bunch of stuff, then you had to miss predicting to back it up. You remembered all the results, you already calculated some of those are just fine. Like if you read a book and you misunderstand paragraph, your understanding of the next paragraph sometimes is invariants that I'm understanding. Sometimes it depends on it </s> |
<s>[INST] How much is fast they I of course is free. [/INST] Everything we do is free. We have no revenue sources of any kind, it's just a service to the community </s><s>[INST] years ST okay. Once a person understands the basics, trains a bunch of models. If we look at the scale of years, what advice do you have for someone wanting to eventually become an expert? [/INST] Train lots of models, specifically trained lots of bottles in your domain area. So an expert what? Right. We don't need more expert. Like uh, create slightly evolutionary research in areas that everybody's studying. We need experts at using deep learning to diagnose </s><s>[INST] malaria. [/INST] Or we need experts at using deep learning to analyze language to study media bias or we need experts in, um, um, analyzing fisheries to identify problem areas in the ocean. You know, that that's, that's what we need. So like become the expert in your passion area and this is a tour which you can use just about anything and you'll be able to do that thing better than other people, particularly by combining it with your passion and domain expertise. </s><s>[INST] So that's really interesting. Even if you do want to innovate on transfer learning or active learning, your thought is it means one I certainly share is you also need to find a domain or dataset that you actually really care [/INST] for. If you're not working on a real problem that you understand? How do you know, if you're doing it any good, how do you know, if your results are good, how do you know, if you're getting bad results, why you're getting bad results? Is it a problem with the data? Like how do you know, you're doing anything useful? Yeah. The only to me the only really interesting research is not the only, but the vast majority of interesting research is like try and solve an actual problem and solve it really well. </s><s>[INST] So both understanding sufficient tools on the deep learning side and becoming a domain expert in particular domain, really things within reach before anybody. [/INST] Yeah, I mean to me, I would compare it to like studying self driving cars, having never looked at the car or been in a car or turn the car on, which is like the way it is for a lot of people, they'll study some academic data set where they literally have no idea about that by </s><s>[INST] the way, I'm not sure how familiar with autonomous vehicles, but that is literally you describe a large percentage of robotics folks working self driving cars as they actually haven't consider driving. They haven't actually looked that we're driving, looks like they haven't driven it [/INST] a problem because you know, when you've actually driven, you know, like these are the things that happened to me when I was driving </s><s>[INST] it, there's nothing that beats the real world examples of just experiencing them. You've created many successful startups, what does it take to create a successful startup? [/INST] Same thing is becoming successful, deep learning practitioner, which is not getting up so you can run out of money or run out of time or run out of something, you know, but if you keep costs super low and try and save up some money beforehand. So you can afford to have some time then just sticking with it is one important thing. Doing something you understand and care about is important by something. I don't mean the biggest problem I see with declining people is they do a PhD in deep learning and then they try and commercialize their PhD is a waste of time because that doesn't solve an actual problem. You picked your PhD topic because it was an interesting kind of engineering or math or research exercise. But if you've actually spent time as a recruiter and you know that most of your time was spent sifting through resumes and you know that most of the time you're just looking for certain kinds of things and you can try doing that with a model for a few minutes and see whether it's something which the model is going to do as well as you could, then you're on the right track to creating a start up and then I think just being just be pragmatic and try and stay away from venture capital money as long as possible, preferably forever. </s><s>[INST] So yeah, at that point, do you venture capital. So did you were able to successfully Iran startups with self funded? [/INST] So my first two was self funded and that was the right way to do it. </s><s>[INST] That's scary. [/INST] No species startups are much more scary because you have these people on your back who do this all the time and you have done it for years telling you grow, grow, grow, grow and I don't, I don't care if you fail. They only care if you don't grow fast enough. So that's scary. Where else doing the ones myself well with partners who were friends. It's nice because like we just went along at a pace that made sense and were able to build it to something which was big enough that we never had to work again. But it was not big enough that any VC would, I think it was impressive and that was enough for us to be excited, you know. Uh, so I thought that so much better way to do things that most people </s><s>[INST] generally speaking that for yourself. But how do you make money during that process? Do you cut into savings? So [/INST] Yeah. So for, so I started fast male and optimal decisions at the same time in 1999 With two different friends and for fast mail I guess I spent $70 a month on the server. And when the server ran out of space, I put a payments button on the front page and said if you want more than 10 mega space you have to pay $10 a yeah. </s> |
<s>[INST] keep my ties with meta, but I am seriously going for the AGI side of things where... And it's actually a really interesting point because a lot of the machine learning, the AI community is quite large, but really, basically almost everybody has taken the same trajectory through life in that community. And it's so interesting to have somebody like you, which are with a fundamentally different trajectory. And that's where the big solutions can come because there is a kind of silo and it is a bunch of people kind of following the same [/INST] kind of set of ideas. And I was really worried that I didn't want to come off as an arrogant outsider for things where I have all the respect in the world for the work. It's been a miracle decade. We're in the midst of a scientific revolution happening now. And everybody doing this is... These are the Einsteins and Bohrs and whatevers of our modern era. And I was really happy to see that the people that I sat down and talked with, everybody does seem to really be quite great about just happy to talk about things, willing to acknowledge that we don't know what we're doing. We're figuring it out as we go along. And I mean, I've got a huge debt on this where this all really started for me because Sam Altman basically tried to recruit me to open AI. And it was at a point when I didn't know anything about what was really going on in machine learning. And in fact, it's funny how the first time you reached out to me, it's like four years </s><s>[INST] ago for your AI pod. Yeah, for people who are listening to this should know that, first of all, obviously, I've been a huge fan of yours for the longest time, but we've agreed to talk like, yeah, like four years ago, back when this was called the artificial intelligence podcast. [/INST] We wanted to do a thing and you said yes. And then I said, I said, it's like, I don't know anything about modern AI. That's right. I said, I could kind of take an angle on machine perception because I'm doing a lot of that with the sensors and the virtual reality, but we could probably </s><s>[INST] find something to talk about. That's right. And that's where one did Sam talk to you about [/INST] open AI around the same time. No, it was a bit after that. So I had done the most basic work. I had kind of done the neural networks from scratch where I had gone and written it all in C just to make sure I understood backpropagation at the lowest level and my nuts and bolts approach. But after Sam approached me, I was flattering to think that he thought that I could be useful at open AI, largely for kind of like systems optimization sorts of things. I am without being an expert, but I asked Ilya Sutskever to give me a reading list and he gave me a binder full of all the papers that like, okay, these are the important things. If you really read and understand all of these, you'll know, like 80% of what most of the, you know, the machine language researchers work on. And, I went through and I read all those papers multiple times and highlighted them and went through and kind of figured the things out there and then started branching out into my own sets of research on things. And I actually started writing my own experiments and doing kind of figuring out, finding out what I don't know, what the limits of my knowledge are and starting to get some of my angles of attack on things, the things that I think are a little bit different from what people are doing, and I've had a couple years now, like two years since I kind of left the full-time position at META, and now I've kind of pulled the trigger and said, I'm going to get serious about it, but some of my lessons all the way back to Armadillo Aerospace about how I know I need to be more committed to this, where there is that it's both a freedom and a cost in some ways when you know that you're wealthy enough to say, it's like this doesn't really mean anything. I can spend $1 million a year for the rest of my life, and it doesn't mean anything, it's fine. But that is an opportunity to just kind of meander, and I could see that in myself when I'm doing some things. It's like, oh, this is a kind of interesting, curious thing. Let's look at this for a little while. Let's look at that. It's not really bearing down on the problem. So there's a few things that I've done that are kind of tactics for myself to make me more effective. One thing I noticed I was not doing well is I had a Google Cloud account to get GPUs there. I was finding I was very rarely doing that for no good psychological reasons where I'm like, oh, I can always think of something to do other than to spin up instances and run an experiment. I can keep working on my local titans or something. But it was really stupid. It was not a lot of money. I should have been running more experiments there. So I thought to myself, well, I'm going to go buy a quarter million dollar DGX station. I'm going to just sit it right there. And it's going to mock me if I'm not using it. If the fans aren't running on that thing, I'm not properly utilizing it. And that's been helpful. I've done a lot more experiments since then. It's been interesting where I thought I'd be doing all this low-level NVLink optimized stuff, but 90% of what I do is just spin up four instances of an </s> |
<s>[INST] in terms of ability to reason is getting quite incredible. There's a lot of demonstrations of it being able to explain jokes so which is not necessarily being able to generate humor yet, but able to explain why something is funny. So there's uh like puns and all those kinds of things, there's there's good benchmarks for that. But you know, if you tell a joke, there's a lot of unspoken stuff that we figure out in our head and it clicks and we understand that it's funny ai is not able to do that, but it's not able to generate the joke yet, as far as I've seen, [/INST] I would say that, I mean just in my experience, I would say that it does because just cause the dude Z is literally, I'll give you another weird example. It's writing a diary Of mine from my childhood. That is not accurate. It's only partially accurate based on the stuff that it can pick up about my life from the age of like 15 of which there isn't much, but I guess we're not, I don't know that what we are, we're laughing our asses off, what the what dudes are you saying? </s><s>[INST] Well, I would say you're laughing, we're laughing our asses off at the collaboration between the human and and the machine [/INST] there because </s><s>[INST] it's it's basically introducing absurdity and into the equation and the kind of absurdity that would together with you create like hilarious stuff, but like on its own, I guess it is in some way writing material for you that's funny, [/INST] but it's very specific </s><s>[INST] to you. It can't do stand up on its own, I guess is what I'm saying. [/INST] That's a good point and that would be terrifying to see an Ai stand up that can actually read a room come up with jokes that could complete that illusion for an audience, but I hear what you're saying, that it needs to be a confluence of both of those elements. And then as you said, like it kind of is it is it is it's kind of, even though it's just for us, and I guess this is I hadn't really thought about this up until right now that in that this company approached us and was like, here's this ai and it's it's it's a podcast ai it's like it chose chad and I, for the reasons that I told you, you know, it's like, here's two guys that do the podcast stuff, they're actually good friends and it knows what's gonna make us laugh, but what is humor uh to uh when it reaches its audience? But the kind of stuff that makes other people laugh at mad tv all we were doing was it was a group of actors and writers and and writer actors and vice versa and um uh who were at at its best, that show was a group of people making each other laugh, you know? And then because we didn't have the internet, we didn't have the feed, the immediate feedback, we had a message board or something. We had emails at the very beginning which check this out people if you have a question or comment mad tv at whatever. Uh and we would get the emails on a monday morning and they would be in a binder or two like this and they would make their way around the </s><s>[INST] office and he's got [/INST] the emails, oh, they're in brian's office, </s><s>[INST] like, like your, your, this is opinions from people about different things. [/INST] The emails that, yeah, the people like literally just writing mad tv. Well, the ones I remember most vividly, yeah, we're fans saying uh </s><s>[INST] you suck, [/INST] yeah, you suck. Like a lot of that when I first started the show for real, you know, because it's a new and you're a new person. It's like, who's this fat bastard? I </s><s>[INST] feel like it that's a [/INST] Good point. Yeah, when you're reading it off of paper and you can literally crunch it up in your hand. But also, it was like, um you know, I would like to see insert weird idea from some 14 year old, you know, I want to see Stewart do this and swan that and but it was, it's a, </s><s>[INST] it's a kind of dude see, but human, [/INST] it was, yeah, it was a very shitty dudes in a loosely finder. But the, the, the thing about the show was we're trying to make each other laugh and do Dzi has found chad and I who we make each other laugh, but it's joined in and it's listen, when I finished doing TMP uh </s><s>[INST] 10 minutes of [/INST] The 10 minute podcast. I didn't really know what I wanted to do in the podcast space and this thing found me and it is genuinely cracking me up anyway, I've said enough about that, but but I do think that it's figured something out with me. It's a really </s> |
<s>[INST] from this place, from this atom to this atom. Just a quick pause, cuz there's a lot of varied people that listen to this. Just to clarify, cuz I did a poll actually. What do you think a graph is a long time ago? And it's kind of funny, how few people know the term graph outside of computer science. So it's- Let's call it network, I think that's- Let's call it a network is better. But every time, I like the word graph though, so let's just say that graph We'll use terms nodes and edges maybe, and it's just nodes represent some abstract entity, and then the edges represent relationships between those entities. Right, exactly. So that's what graphs, sorry, so there you go. So that's the basic structure. [/INST] That's a MT, let's call it a network. I think that's it. Let's call it a network. Right, exactly. That is the simplest case of a basic structure. Actually, it tends to be better to think about hypergraphs. So a hypergraph is just, instead of saying there are connections between pairs of things, we say there are connections between any number of things. So there might be ternary edges. So instead of just having two points are connected by an edge, you say three points are all associated with a hyperedge, are all connected by a hyperedge. just at some level that's a detail. It's a detail that happens to make the, for me, in the history of this project, the realization that you could do things that way broke out of certain kinds of arbitrariness that I felt that there was in the model before I had seen how this </s><s>[INST] worked. All a hypergraph can be mapped to a graph. It's just a convenient representation, [/INST] mathematical speaking. Right. That's correct. Okay. So the first question, the first idea of these models of ours is space is made of these connected sort of atoms of space. The next idea is space is all there is. There's nothing except for this space. So in traditional ideas in physics, people have said there's space, it's kind of a background, and then there's matter, all these particles, electrons, all these other things which exist in space. Right. But in this model, one of the key ideas is there's nothing except space. So in other words, everything that has, that exists in the universe is a feature of this hypergraph. So how can that possibly be? Well, the way that works is that there are certain structures in this hypergraph where you say that little twisty knotted thing, we don't know exactly how this works yet, but we have sort of idea about how it works mathematically. This sort of twisted knotted thing, that's the core of an electron. This thing over there that has this different form, </s><s>[INST] that's something else. So the different peculiarities of the structure of this graph are the very things that we think of as the particles inside the space, but in fact, it's just the property of this space. Mind blowing, first of all. It's mind blowing, [/INST] and we'll probably talk in its simplicity and beauty. In fact, it's just the property of this </s><s>[INST] space. Yes, I think it's very beautiful. But that's space, and then there's another concept we didn't really kind of mention, but you think it of computation as a transformation. [/INST] Let's talk about time in a second. Let's just, I mean, on the subject of space, the question of kind of what, there's this idea. There is this hypergraph, it represents space, and it represents everything that's in space. The features of that hypergraph, you can say certain features and this part, we do know certain features of the hypergraph represent the presence of energy, for example, or the presence of mass or momentum. We know what the features of the hypergraph that represent those things are, but it's all just the same hypergraph. One thing you might ask is if you just look at this hypergraph. You say, and we're going to talk about what the hypergraph does. But if you say how much of what's going on in this hypergraph is things we know and care about, like particles and atoms and electrons and all this kind of thing, and how much is just the background of space? So it turns out, so far as in one rough estimate of this, everything that we care about in the universe is only one part in 10 to the 120 of what's actually going on. The vast majority of what's happening is purely things that maintain the structure of space. In other words, the things that are the features of space that are the things that we consider notable, like the presence of particles and so on, that's a tiny little piece of froth on the top of all this activity that mostly is just intended to, you know, mostly I can't say intended, there's no intention here </s><s>[INST] that just maintains the structure of space. Let me load that in. It just makes me feel so good as a human being to be the froth on the one in a 10 to the 120 or something. And also just humbling how, in this mathematical framework, how much work needs to be done on the infrastructure [/INST] of our universe. To maintain the infrastructure of our universe is a lot of work. We are merely writing little tiny things on top of that infrastructure. But, you know, you were just starting to talk a little bit about space that represents all the stuff that's in the universe. The question is, what does that stuff do? And for that, we have to start talking about time and what is time and so on. And, you know, one of the basic idea of this model is time is the progression of computation. So in other words, we have a structure of space and there is a rule that says how that structure of space will change. And it's the application, the repeated application of that rule that defines the progress of time. </s> |
<s>[INST] So, but what do you think of the Tesla data engine approach? Which to me is an active learning task is kind of fascinating. Is breaking it down into these multiple tasks and mining their data constantly for like edge cases for these different tasks. But the [/INST] tasks themselves are not being learned. This is feature engineering. </s><s>[INST] Yeah, I mean it's it's uh it's a higher abstraction level of feature engineering for the different tasks. Task engineering in a sense, it's [/INST] slightly better feature engineering but it's still fundamentally is feature engineering and anything about the history of AI has taught us anything. It's that feature engineering approaches will always be replaced and lose to end to end. Now to be fair, I cannot really make promises on timelines. But I can say that when you look at the code for stock fish and the code for α 01 is a lot shorter than the other. A lot more elegant required a lot less program or ours too. Right. </s><s>[INST] Yeah. But there was a lot more murder of bad Agents on the Alpha zero side by murder. I mean uh agents that played a game and failed miserably. Oh, in simulation that failure is less costly in in real world. Do [/INST] You mean in practice like α zero has lost games miserably? </s><s>[INST] No, I haven't seen that. No, but I know, but the requirement for Alpha zero is to be able to like evolution. Human evolution, not human evolution. Biological evolution of life on earth from the origin of life has murdered trillions upon trillions of organisms on the path to us humans. So the question is, can we stitch together a human like object without having to go through the entire process of evolution? [/INST] Well, no, but do the evolution of simulation. </s><s>[INST] Yes, this is the question, can we simulate. So do you have a sense that's possible to simulate some U [/INST] zero is exactly this. Mu zero is the solution to this. Mu Zero I think is going to look be looked back as the canonical paper. And I don't think deep learning is everything. I think that there's still a bunch of things missing to get there. But mu zero, I think is going to be looked back as the kind of cornerstone paper um, of this whole deep learning era and mu zero is the solution to self driving cars. You have to make a few tweaks to it. But mu zero does effectively that it does those rollouts and those murdering in learned simulator and learned dynamics model. </s><s>[INST] It's interesting. It doesn't get enough. Love that. I was blown away [/INST] when I was blown away when I read that paper, I'm like, you know, okay, I've always had a comma, I'm gonna say I'm gonna wait for the solution to self driving cars to come along This hero sight. It's μ zero. </s><s>[INST] Yeah. So, uh, [/INST] sit back and let the winning role in. </s><s>[INST] So your sense just to elaborate a little bit to linger on the topic, Your senses, neural networks will solve driving like we don't need anything else. [/INST] I think the same way chess was maybe the chest and maybe google are the pinnacle of like search algorithms and things that look kind of like a star. Um, the pinnacle of this era is going to be self driving cars. </s><s>[INST] Yeah. But on the, on the path that you have to deliver products and it's possible that the path to full, self driving cars will take decades. I doubt it. How long would you put on it? Like, what do we, you're chasing it? Tesla's chasing it. What are we talking about? five years, 10 years, 15 years [/INST] In the 2020s </s><s>[INST] In the 2020's [/INST] The later part of the 2020s. </s><s>[INST] Well, the neural network that would be nice to see and on the path to that you're delivering products which is a nice L. Two system, that's what Tesla is doing. A nice L. Two system [/INST] just gets better every time. L two. The only difference between L. Two and the other levels is who takes liability and I'm not a liability. God only take liability level two forever. Mhm. </s><s>[INST] Now on that little transition, I mean, how do you make the transition work? Is this where driver sensing comes in? Like how do you make the because you said 100 miles like is there some sort of human factor psychology thing where people start to over trust the system, all those kinds of effects once it gets better and better and better and better, they get lazier and lazier and lazier. Is that like how do you get that transition? Right [/INST] first off our monitoring is already adaptive monitoring is already seen, adaptive </s><s>[INST] driver monitoring is just the camera that's looking at the driver. You have an infrared camera in the [/INST] our policy for how we enforce the driver monitoring is seen adaptive. What's that mean? Well, for example, in one of the extreme cases, um, if you if the car is not moving, we do not actively enforced driver monitor. All right. Um if you are going through a uh like a 45 mile an hour road with lights um and stop signs and potentially pedestrians, we enforce a very tight driver monitoring policy. If you are alone on a perfectly straight highway. Um and this is, it's all machine learning, none of that is hand coded actually that stopped his hand coded, but </s> |
<s>[INST] with open ai all the exciting projects that you're working on, What is 2019? [/INST] Hold for you massive scale scale, I will put an asterisk on that and just say you know, I think that it's about ideas plus scale, you need both. </s><s>[INST] So that's a really [/INST] good point. </s><s>[INST] So the question in terms of ideas, you have a lot of projects that are exploring different areas of intelligence. And uh the question is when you when you think of scale, do you think about growing scale those individual projects or do you think about adding new projects and uh sorry too, if you're thinking about adding new projects or if you look at the past, what's the process of coming up with new project of new ideas? [/INST] Uh So we really have a life cycle of project here. Uh So we start with a few people just working on a small scale idea and language is actually very good example of this. That is really you know, one person here who was pushing on language for a long time. I mean then you get signs of life. Right? And so this is like let's say, you know, with with the original GPT We had something that was interesting and we said Okay, it's time to scale this, right, it's time to put more people on it, put more computational resources behind it. And uh and then we just kind of keep pushing and keep pushing. And the end state is something that looks like Dota or robotics where you have a large team of you know, 10 or 15 people that are running things at very large scale and that you're able to really have material engineering. Uh and and uh and you know, sort of machine learning science coming together to make systems that work uh and get material results that just would've been impossible otherwise. Um So we do that whole lifecycle, we've done it a number of times. Uh you know, typically end to end, it's probably 22 years or so to do it. You know, the organization's been around for three years. So maybe we'll find that we also have longer life cycle projects. But you know, we we uh we'll work up to those. We have so so 11 team that we were actually just starting, Ellie and I are are kicking off a new team called the reasoning team and that this is to really try to tackle. How do you get neural networks to reason? And uh we think that this will be a long term project and we're very excited about </s><s>[INST] uh in terms of reasoning, super exciting topic. What do you what kind of benchmarks? What kind of tests of reasoning do you envision? What what would if you set back uh whatever drink and you would be impressed that this system is able to do something? What would that look [/INST] like they're improving? </s><s>[INST] They're improving. So some kind of logic and especially mathematical logic? [/INST] I think so. Right. I think that there's there's there's kind of other problems that are dual to theorem proving in particular. Um You know, you think about programming, I think about even like security analysis of code um that these all kind of capture the same sorts of core reasoning and being able to do some out of distribution generalization. </s><s>[INST] It would be quite exciting if open ai reasoning team was able to prove that P equals Np. That'll be very nice. [/INST] It would be very very very exciting. Especially if it turns out that P equals np. That'll be interesting to </s><s>[INST] uh just, it would be uh ironic and humorous. Uh So what problem stands out to you is uh the most exciting and challenging, impactful to the work for us as a community in general. And for open ai this year, you mentioned reasoning. I think that's that's a heck of a [/INST] Problem. Yeah, so I think reasoning is an important one. I think it's gonna be hard to get good results in 2019. Um you know, again, just like we think about the life cycle takes time. Um I think for 2019 language modeling seems to be kind of on that ramp, right? It's at the point that we have a technique that works. We want to scale 100 X 1000 ducks. See what happens </s><s>[INST] awesome. Do you think we're living in a simulation? [/INST] Uh I think it's I think it's hard to have a real opinion about it. You know, it's actually interesting. I separate out things that I think can have like, you know, yield materially different predictions about the world from ones that are just kind of fun to speculate about. And I kind of view simulation, it's more like is there a flying teapot between mars and jupiter? Like maybe, but it's a little bit hard to know what that would mean for my life. </s><s>[INST] So there is something actionable. So some of the best work opening has done is in the field of reinforcement learning and uh some of the success of reinforcement learning come from being able to simulate the problem and trying to solve. So do you have a hope for reinforcement, for the future of reinforcement learning and for the future of simulation? Like whether we're talking about autonomous vehicles or any kind of system. Do you see that scaling or we'll be able to simulate systems and hence be able to create a simulator that echoes are real world and proving once and for all, even though you're denying it that we're living in a simulation. [/INST] I feel it's two separate questions. Right? So, you know, kind of the core there of like can we use simulation for self driving cars. Um take a look at our robotic system doctor. All right. That was trained in simulation using the data system. In fact, and it transfers to a physical robot. And I think everyone looks at our data system that like, okay, it's just a game. How are you ever gonna escape to the real world? And the answer is, well, we did it with the physical robot that no one can program. And so, I think the answer is simulation goes a lot further than you think if you apply the right techniques to it. Um, Now there's a question of, you know, are the being in that simulation gonna gonna wake up and and have consciousness? Um, I think that one seems a lot a lot harder to to again, reason about, I think that, you know, you really should think about like, where exactly just human consciousness come from in our own self awareness. And, you know, is it just that like once you have like a complicated enough neural net, you have to worry about the agents feeling pain. Um, and uh, you know, I think there's like, interesting speculation to to do there, but, uh, but, you know, again, I think it's a little bit hard to know for sure. </s> |
<s>[INST] Uh Yes. So let me ask a silly sounding ridiculous sounding but probably extremely important question sex in space including intercourse, conception, procreation, birth like being a parent like raising the baby. So basically from birth well from the before birth part um like the birds and the bees and stuff and then the whole thing how complicated is that? I remember looking at the thank you. I remember looking at this exact Wikipedia page actually and it's I remember being uh the Wikipedia pages. Sex in space. And fascinating how difficult of an engineering problem. The whole thing is is that something you think about to how to have generations of humans self self replicating [/INST] these </s><s>[INST] societies essentially. I mean I guess with Micro like if you solve the gravity problem you solve a lot of these problems. [/INST] That's the hope. Yeah it's like the central challenge of microgravity to human reproduction. But we do host a workshop every year and beyond the cradle which is the space event that we rented M. I. T. And we always do want on pregnancy in space or motherhood or raising Children in space because there are huge questions um there have been a few mammal studies that have looked at reproduction in space. But there are still really major questions about how does it work? How does the fetus evolve in microgravity if you were pregnant in space? And I think the near term answer is just gonna be we need to be able to give humans a one g environment for that phase of our development. </s><s>[INST] Yeah. So there's some studies on mice in microgravity and it's interesting like I think the mice like one of them the mice weren't able to walk or like their their understanding of physics I guess is all for something like that. [/INST] Yeah. The mental model when you're really young and you're kind of getting your mental model of physics. We do think that that would change um kids abilities to if they were born in microgravity, their ability to have that intuition around an Earth based one G environment might be missing because a lot of that is really crystallized in early development, early childhood development. So that makes sense that they would see that in mice. </s><s>[INST] Yeah. So what about life when we are choose to park our vehicles on another planet on the moon? But let's go to mars first of all, does that excite you humans going to mars like stepping foot on mars and when do you think it will happen? It [/INST] does excite me. I think visionaries like Elon are working to make that happen in terms of building the road to space. We are really excited about building out the human lived experience of space once you get there. So how are you going to grow your food? What is your habitat gonna look like? I think it's profoundly exciting, but I do think that there's a little bit of a misunderstanding of MArs anywhere in the near future being anything like a replacement for earth. So it is good for humanity to have these other pockets of our civilization that can expand out beyond earth. But MArs is not in its current state, a good home for humanity. Too many prickly rates in the soil. You can't use that soil to grow crops. Atmosphere is too thin, certainly can't breathe it, but it's also just really thin compared to our atmosphere, a lot of different challenges that would have to be fundamentally changed on that planet to make it a good home for a large human civilization. </s><s>[INST] How does a large civilization of humans get built on MArs? And what do you think? Um Where do you think it gets started being difficult? So, can you have a small base of like 10 people essentially kind of like the international space station kind of situation? And then can you get it to 100 to 1000 to a million. Are there some interesting challenges there that worry you saying that Mars is just not a good backup at this time for earth? [/INST] I think small outpost. Absolutely like McMurdo. Right, so we have these models of really extreme environments on earth in Antarctica for example, where humans have been able to go and make a sustainable settlement. McMurdo style life on MArs probably feasible in the twenties thirties. So we want to send the first human missions to MArs and maybe as early as the end of this decade, more likely early twenties thirties moving anywhere beyond that in terms of a place where like an entire human life would be lived, where it's not just you go for a three month deployment and you come back. That is actually the big challenge line is just saying is there enough um technological sophistication that can be brought that far out into space? If you imagine your electronics break, there's no radio shack. This dates me a little bit that my mind jumps to radioshack but there's no, you know, there's no supply chains on MArs that can supply the level of technological sophistication for all the products that we rely on on day to day life. So you'd be going back to actually a very simple existence more like pioneer life out west in the story of the U. S. For example. And I think that the future of larger scale gatherings of humans in orbit or so in space is actually gonna be in microgravity floating space cities, not so much trying to establish settlements on the surface. </s><s>[INST] So you think sort of a significant engineering investment in terms of our efforts and money should be on large um spaceships that perhaps are doing this kind of self assembly, all these kinds of things and doing an orbit, maybe building a giant donut around the planet over time. [/INST] Yeah, that is the goal. And I think the current political climate is such that you can't get the trillion dollar investment to build, to start from scratch and build the sci fi mega structure. But if you can build it in fits and starts in a little different pieces, which is another advantage of self assembly, it's much more like how nature works. So it's biomimicry inspired way for humanity to scale out in space, </s> |
<s>[INST] Also the ceo of the company on. Yeah, if we get her last name starts with an F. I think it's Fernald follow her on instagram or wherever else she's active because she happens to be a brilliant chef and just has a scientific view of agriculture and food in general, which I find fascinating and inspiring Anyway. You can order belcampo sustainably raised meats to be delivered straight to your door using code lex at belcampo dot com slash lex For 20% of the first time customers that's collects at belcampo dot com slash lex. Trust me the extra bit of cost is worth it. And now here's my conversation with Jim Keller. What's the value and effectiveness of theory versus engineering this dichotomy in building good software or hardware systems? [/INST] Well, it's good designs, both I guess that's pretty obvious by the engineering. Do you mean you know, reduction of practice of known methods and then science is the pursuit of discovering things that people don't understand or solving unknown problems, </s><s>[INST] definitions are interesting here, but I was thinking more in theory constructing models that kind of generalize about how things work and engineering is actually building stuff. The pragmatic like, okay we have these nice models but how do we actually get things to work? Maybe economics is a nice example like economists have all these models of how the economy works and how different policies will have an effect. But then there's the actual, okay let's call it engineering of actually deploying the policies. [/INST] So computer design is almost all engineering and reduction of practicing known assets now, because of the complexity of the computers we built. You know, you could think you're, well, we'll just go write some code and then we'll verify it and we'll put it together and then you find out that the combination of all that stuff is complicated and then you have to be inventive to figure out how to do it, right? So that's that's definitely happens a lot. And then every so often some big idea happens, but it might be one person </s><s>[INST] and that idea is in what, in the space of engineering or is it in the space? [/INST] I'll give you example, so one of the limits of computer performances, branch prediction. So, and there's a there's a whole bunch of ideas about how good you could predict the branch and people said there's a limit to it and thats methodic curve. And somebody came up with a better way to do branch prediction. That was a lot better. And he published a paper on it and every computer in the world now uses it. And it was one idea. So the the engineers who build branch prediction hardware, we're happy to drop the one kind of training array and put it in another one. So it was it was a real idea </s><s>[INST] and parents prediction is, is one of the key problems underlying all of sort of the lowest level of software. It boils down to branch prediction [/INST] boils down uncertainty. Computers are limited by a single thread, computers limited by two things, the predictability of the path of the branches and the predictability of the locality of data. So we have predictors that now predict both of those pretty well. Yeah. So memories, you know, a couple 100 cycles away. Local cache, just couple cycles away. When you're executing fast, virtually all the data has to be in the local cache. So simple program says, you know, add one to every element in array. It's really easy to see what the stream of data will be, but you might have a more complicated program that, you know, so it's going to get an element of this array, look at something and make a decision, go get another element. It's kind of random and you can think that's really unpredictable. And then you make this big predictor that looks at this kind of pattern, then you realize, well, if you get this data and this data, then you probably want that one. And if you get this one and this one and this one, you probably want that one. </s><s>[INST] And is that theory, is that engineering? Like the paper that was written, was it uh was asymptotic kind of kind of discussion? Or is it more like, here's a hack that works well, [/INST] um it's a little bit of both. Like there's information theory and I think somewhere </s><s>[INST] just it's actually trying to prove, [/INST] but once, once, you know, the method implementing it is an engineering problem. Now there's a flip side of this which is in a big design team, what percentage of people think they're there there their their plan or their life's work is engineering versus design, inventing things. So lots of companies will reward you for filing patents. Some many big companies get stuck because they get promoted, You have to come up with something new. And then what happens is everybody is trying to do some random new thing, 99% of which doesn't matter. And the basics get neglected and or they get too, there's a dichotomy. They think like the cell library and the basic cad tools, you know, our basic, you know, software validation methods, that simple stuff. You know, they want to work on the exciting stuff and then they spend lots of time trying to figure out how to pad in something and that's mostly useless. </s><s>[INST] But the breakthroughs on simple stuff, [/INST] no, no. You know, you have to do the simple stuff really well if you're building a building out of bricks, you want great bricks. So you go to two places to sell bricks. So one guy says, yeah, they're over there and ugly pile. And the other guy is like lovingly tells you about the 50 kinds of bricks and how hard they are and how beautiful they are, you know, square they are. You know which one you can buy bricks for them, which is going to make a better house. </s><s>[INST] So you're talking about the craftsman, the person who understands bricks, who loves breaks, who loves [/INST] them, right? That's a good word. Good engineering is great craftsmanship. And when you start thinking engineering is about invention and set up a system that rewards invention, The craftsmanship gets neglected. </s><s>[INST] Okay, so maybe one perspective is the theory, the science overemphasizes invention and engineering emphasizes craftsmanship and therefore like, so if you, it doesn't matter what you do theory, [/INST] everybody does, like read the tech racks are always talking about some breakthrough or intervention innovation and everybody thinks that's the most important thing. But the number of innovative ideas is actually relatively low. We need them write an innovation creates a whole new opportunity. Like when when some guy invented the internet, right? Like that was a big thing. The million people that wrote software against that, we're mostly doing engineering. Software writing the elaboration of that idea. It was huge. </s><s>[INST] I don't know if, you know, Brendan, I, he wrote Javascript in 10 days. That's an interesting story. It makes me wonder. And it was, you know, famously for many years considered to be a pretty crappy programming language still is perhaps has been improving sort of consistently. But the interesting thing about that guys, you know, he doesn't get any awards. You don't get a Nobel prize or fields metal or uh, [/INST] a crappy piece of that software code that </s><s>[INST] that is currently the number one programming language in the world and runs now increasingly running the backhand of the internet. Does he, does he [/INST] know why everybody uses it? Like that would be an interesting thing. Was it the right thing at the right time? Because like when stuff like javascript came out, like there was a move from, you know, writing C programs and C plus plus two, let's call what they call managed code frameworks where you write simple code, it might be interpreted, it has lots of libraries, productivity is high and you don't have to be an expert. So, you know, Java was supposed to solve all the world's problems. It was complicated. Javascript came out, you know, after a bunch of other scripting languages. I'm not an expert on it. But what was it the right thing at the right time? Or was there something, you know, clever because he wasn't the only one, </s><s>[INST] There's a few [/INST] elements maybe if you figured out what it was. No, then you get a prize like that destructive, you know, baby. It's probably hasn't defined this where it just needs a good promoter. </s><s>[INST] Well, I think there was a bunch of blog post written about it, which is like wrong is right, which is like doing the crappy thing fast. Just like hacking together thing that answers some of the needs and then iterating over time listening to developers, like listening to people who actually use the thing, this is something you can do more in software. Uh but the right time, like you have to sense, you have to have a good instinct of one is the right time for the right tool and make it super simple and just get it out there. The problem is this is true with hardware, this is less true of software. Is this backward compatibility that just drags behind you? There's, you know, as you try to fix all the mistakes of the past. But the timing, it was good. There's something about that. It wasn't accidental. You have to like give yourself over to the you have to have this like broad sense of what's needed now, both scientifically and like the community and just like this, it was obvious that uh there was no the interesting thing about javascript is everything that ran in the browser at the time, like java and and I think other like scheme other programming languages, they were all in a separate external container. And then javascript was literally just injected into the web page. That is the dumbest possible thing running in the same thread as everything else. And like uh it was inserted as a comment. So javascript code is inserted as a comment in the html code. And it was I mean there's it's either genius or super dumb, but it's like [/INST] it has no apparatus for like a virtual machine and container, it just executed in the framework, the program is already running and it was </s><s>[INST] cool. And then because something about that accessibility, the ease of its use uh resulted in then developers innovating of how to actually use it. I mean, I don't even know what to make of that, but it does seem to echo across different software, like stories of different software. PHP has the same story. Really crappy language they just took over the world. [/INST] We'll just have a joke that the random length instructions that variable length instructions, that's always one. Even though they're obviously worse, like nobody knows why X-86 is or you'll be the worst architecture, you know, on the planet is one of the most popular ones? </s><s>[INST] Well, I mean isn't isn't that also the story of risk or this is I mean is that simplicity? There's something about simplicity that us in this evolutionary process is valued if it's simple, it gets it spreads faster, it seems like or is that not always [/INST] true? That's not true. Yeah, it could be simple is good, but too simple. It's bad. </s><s>[INST] So why did risk when you think so far? Okay, in the long arc of [/INST] history, we don't know. </s><s>[INST] So who's going to win? What's risk? What's risk and who's going to win in that space in these instruction sets? [/INST] Hey, Ice offers gonna win. But there'll be little computers that run little programs like normal all over the place, but we're going through another transformation. </s><s>[INST] So but you you think instruction sets underneath it all will change. [/INST] Yeah, they evolve slowly. They don't matter very much. They don't matter very much. Okay. I mean the the limits of performance are, you know, predictability of instructions and data. I mean that's the big thing and then the usability of it is some, you know, quality of design, quality of tools, availability, but right now X 86 is proprietary with Intel AMG but they can change it any way they want independently right arm's proprietary to arm and they won't let anybody else change it. So it's like a soul point And risk five is open source so anybody can change it, which is super cool. But that also might mean it gets changed into many random ways that there is no common subset of it that people can use. </s><s>[INST] Do you like open or do you like clothes like if you were to bet all your money on one or the other? Risk five versus no idea. [/INST] It's case dependent. Well, X 86 oddly enough, when Intel first started developing a day licensed like seven people. So it was the open architecture And then they move faster than others and also bought one or 2 of them. But there was seven different people making X 86 because at the time there was 65 02 and Z eighties and you know, 80 86 And you could argue everybody thought Z80 was the better instruction said, But that was pra priority proprietary to one place in the 6800. So there's like 5, 4 or five different microprocessors until went open, got the market share because people felt like they had multiple sources from it. And then over time it narrowed down to two players. </s><s>[INST] So why you as a historian uh why did intel, when for so long with with their processors? I [/INST] mean rate, their process development was great. </s><s>[INST] So it's just looking back to javascript and panic is uh Microsoft and Netscape and all these uh internet browsers. Microsoft won the browser game because they aggressively stole other people's ideas. Like right after they did it. [/INST] You know, I don't know if intel was stealing other people's ideas. They started making way stealing. Just to clarify, they started making RAMs random access memories. And then at the time when the Japanese manufacturers came up, you know, they were getting out competed on that and they pivoted the microprocessors and they made the first, you know, integrated microprocessor grant programs, the uh 40 oh four or something </s><s>[INST] who was behind that pit, That's a hell of a pivot. [/INST] Andy Grove and he was great. </s><s>[INST] That's a hell of a pivot. [/INST] And then they lead semiconductor industry. Like they were just a little company IBM all kinds of big companies had boatloads of money and they out innovated. Everybody </s><s>[INST] out innovative. Okay. Yeah, it's not like marketing, it's [/INST] not their processor designs were pretty good. Um I think the, You know, Core two was probably the first one I thought was great. It was a really fast processor and then house well it was great. </s><s>[INST] What makes a great process during that? [/INST] Oh, if you just look at its performance versus everybody else, it's, you know, the size of it, the usability of it, </s><s>[INST] so it's not specific, some kind of element that makes it beautiful. It's just like literally just raw performance. Is that how you think of bio processors is just like raw performance? Of course, [/INST] It's like a horse race, the fastest one wins. </s><s>[INST] Now you don't care [/INST] how well there's the fastest in an environment, like for years you made the fastest one you could and then people started to have power limits, so and then you made the fastest at the right power point. And then and then when we started doing multi processors, like If you could scale your processor is more than the other guy, you could be 10% faster on a single thread, but you have more threads. so there's lots of variability. And then arm. Okay. Really explored like, you know, they have the a series in the R series in the m series, like a family of processors for all these different design points from like unbelievably small and simple. And so then when you're doing the design, it's sort of like this big palette of Cpus, like they're the only ones with a credible you know top to bottom pallet and </s><s>[INST] what do you mean incredible talk [/INST] about? Well there's people who make micro controllers that are small but they don't have a fast one. There's people make fast processors but don't have a little medium under a small one. </s><s>[INST] Is that hard to do that full palette that seems like [/INST] a lot of different. </s><s>[INST] So what's the difference in the armed folks and and tell in terms of the way they're approaching this problem. [/INST] Well until Almost all the processor designs were you know very custom high end you know for the last 1520 years </s><s>[INST] the fastest force possible in one horses. [/INST] Yeah and and the architecture that are really good. But the company itself was fairly insular to what's going on the industry with cad tools and stuff. And there's this debate about custom designed versus synthesis and how do you approach that? I'd say Intel was slow on the getting to synthesize processors. Arm came in from the bottom and they generated I. P. Which went to all kinds of customers so they had very little say in how the customer implemented their I. P. So arm is super friendly to the synthesis ip environment whereas Intel said we're gonna make this great client chip server chip with our own cattles with our own process with our own you know other supporting I. P. And everything only works with our stuff </s><s>[INST] so is that uh his arm winning the mobile platform space in terms of trust and so in that what you're describing is why they're winning. [/INST] Well they had lots of people doing lots of different experiments so they control the process of architecture and I. P. But they let people put in lots of different ships and there was a lot of variability in what happened there. Whereas intel when they made their mobile their foray into mobile they had one team doing one part Right? So it wasn't 10 experiments. And then their mindset was PC mindset Microsoft Software mindset and that brought a whole bunch of things along that the mobile world, embedded world don't. Do </s><s>[INST] you think it was possible for intel to pivot hard and win the mobile market? That's a hell of a difficult thing to do right? For a huge company to just pivot. I mean it's so interesting too because we'll talk about your current work it's like it's clear that pcs were dominating for several decades like desktop computers and then mobile it's unclear [/INST] it's the leadership question like like Apple under steve jobs when he came back they pivoted multiple times. You know they built ipads and itunes and phones and tablets and great max. Like like who knew computers should be made out of aluminum. Nobody knew that. That's great. It's super fun though steve yeah steve jobs like they pivoted multiple times and uh you know the old intel? They did that multiple times. They made the ramps and processors and processes and </s><s>[INST] I gotta ask this, what was it like working with steve jobs? I [/INST] didn't work with him. </s><s>[INST] Did you interact with [/INST] him twice? I said hi to him twice in the cafeteria. </s><s>[INST] What did you say? [/INST] Hi. He said, hey fellas, he was friendly, he was wandering around and with somebody, he couldn't find the table because the cafeteria was, it was packed and I gave my table but I worked for Mike Culbert who talked to like Mike was the unofficial Ceo of Apple And a brilliant guy. And he worked for Steve for 25 years maybe more. And he talked to steve multiple times a day and he was one of the people who could put up with steve, let's say brilliance and intensity and steve really liked him. And steve trusted Mike to translate the ship, he thought up into engineering products that work. And then Mike ran a group called platform architecture. And I was in that group so many times I'd be sitting with Mike and the phone would ring maybe steve. And Mike would hold the phone like this because steve would be yelling about something or other and he translated and then he would say steve wants us to do this </s><s>[INST] so well, steve a good engineer or now I don't know, [/INST] he was a great idea guy idea person. He's a really good selector for talent </s><s>[INST] Yeah, it seems to be one of the key elements of leadership, right? [/INST] And then he was a really good first principles guy like like somebody say something couldn't be done and he would just think that's obviously wrong, right? But you know maybe it's hard to do, maybe it's expensive to do. Maybe we need different people, you know, there's like a whole bunch of if you want to do something hard maybe it takes time, maybe you have to iterate, there's a whole bunch of things that you can think about but saying it can't be done is stupid, </s><s>[INST] how would you compare? So it seems like Elon musk is more engineering centric but it's also I think he considered himself a designer too has a design mind steve jobs feels like he is much more idea space design, space versus engineering just make it happen like the world should be this way, just figure it out. [/INST] But he used computers, you know, he had computer people talk to them all the time. Like mike was a really good computer guy. He knew computers could do </s><s>[INST] computer meeting computer hardware, like hardware, [/INST] software, all the pieces and then he would I have an idea about what could we do with this next. That was grounded in reality, it wasn't like he was you know just finger painting on the wall and wishing somebody would interpret it like so he had this interesting connection because no he wasn't a computer architecture designer but he had an intuition from the computers, we had to, what could happen and </s><s>[INST] essentially say intuition because it seems like he was pissing off a lot of engineers in his intuition about what Canada can be done. Those like the, what is all these stories about like floppy disks and all that kind of stuff? Like [/INST] yeah, so, and steve the first round like he'd go into a lab and look at what's going on and hate it and, and uh fire people or assembly in the elevator what they're doing for Apple and you know, not be happy when he came back. My impression was is he surrounded himself with this relatively small group of people and didn't really interact outside of that as much. And then the joke was, you see like somebody moving a prototype through the quad with a, with a black blanket over it and that was because it was secret, you know, partly from steve because they don't want steve to see it until it was ready. </s><s>[INST] Yeah. The dynamical johnny Ive and steve is interesting. It's like, you don't wanna, he ruins as many ideas as he generates. Yeah, It's a dangerous kind of line to walk. [/INST] If you have a lot of ideas. Like Gordon Bell was famous for ideas, right? And it wasn't that the percentage of good ideas was way higher than anybody else. It was, he had so many ideas and, and he was also good at talking to people about it and get into filters, right? And you know, seeing through stuff, where's Ellen was like, hey, I want to build rockets. So steve was hire a bunch of Rocket guys and Ellen would go read Rocket Mandel's </s><s>[INST] so yon is a better engineer a sense like or like more uh like a love and passion for the manuals [/INST] and the details, the details, the </s><s>[INST] craftsmanship too. Right, Well I guess you had craftsmanship to a bit of a different kind. What do you make of the just to stand in for just a little longer, what do you make of like the anger and the passion and all that? The firing and the mood swings and the madness, the, you know, being emotional and all that. That's steve. And I guess Ellen too, so what is that uh that a bugger feature? [/INST] The feature? So there's a graph which is uh Y axis, productivity X axis at zero, it's chaos and then finishes complete order. Yeah. Right, so as you go from the, you know the origin as you improve order, you improve productivity and at some point productivity peaks and then it goes back down again, too much order, Nothing can happen. Yes, </s><s>[INST] but the question is, how close to the chaos is [/INST] that? No, no, no. Here's the thing is once you start moving the direction order, the force factor to drive you towards order is unstoppable, </s><s>[INST] oh, [/INST] every organization will move to the place where their productivity is stymied by order. </s><s>[INST] So you need uh [/INST] So the question is, who's the counter for us, Like? Because it also feels really good as you get more organized and productivity goes up. The organization feels they oriented towards it right? To hire more people. They got more guys who can run process, you get bigger, right? And then, inevitably, inevitably, the organization gets captured by the bureaucracy that manages all the processes, alright? And then humans really like that. And so, if you just walk into a room and say, guys love what you're doing, but I need you to have less order. Mhm. If you don't have some force behind that, nothing will happen. </s><s>[INST] I can't tell you on how many levels that's profound. So, [/INST] so that's why I say it's a future. Now, could you be nicer about it? I don't know, I don't know any good examples of being nicer about it. Well, the funny thing is to get stuff done, You need people who can manage stuff and manage people because humans are complicated. They need lots of care and feeding that you need to tell them they look nice and they're doing good stuff and pat them on the back, </s><s>[INST] right? I don't know, You tell me, is that does that needed humans need [/INST] that I had a friend who started a magic group and he said I figured it out, you have to praise them before they do anything I was waiting till they were done. And they're always mad at me now? I tell them what a great job you're doing while they're doing. But then you get stuck in that trap because then when you're not doing something, how do you confront these people? </s><s>[INST] I think a lot of people that had trauma in their childhood to disagree with you, successful people that you just first do the rough stuff and then be nice later. I don't know. [/INST] Okay. But you know, engineering companies are full of adults who had all kinds of ranch of childhoods. You know, people had okay childhood. </s><s>[INST] Well, I don't know if [/INST] lots of people only work for praise, which is weird. </s><s>[INST] You mean like everybody, [/INST] I'm not that interested in this, but </s><s>[INST] uh, well you're, you're, you're probably looking for somebody's approval. Mm Even still, [/INST] yeah, maybe </s><s>[INST] I should think about that. Maybe somebody who's no longer with this kind of thing. I don't know. [/INST] I used to call it my dad and tell him what I was doing. He was, he was very excited about engineering and </s><s>[INST] stuff. You've got his approval. [/INST] Uh, yeah, a lot. I was lucky. Like he, he decided I was smart and unusual as a kid and that was okay when I was really young. So when I directed poorly in school I was dyslexic, I didn't read until I was third or fourth grade and they didn't care. My parents were like, oh, he'll be fine. So that was that was cool. </s><s>[INST] Is he still with us? You miss him. [/INST] Sure. Yeah. Parkinson's and then cancer. His last 10 years were tough and I killed him. Killing a man. Like that's hard </s><s>[INST] the mind. [/INST] Well it's pretty good. Um Parkinson's caused a slow dementia and uh, the chemotherapy I think accelerated it. But it was like hallucinogenic dementia. So he was clever and funny and interesting. And was it was pretty unusual. </s><s>[INST] Do you remember conversations uh from that time? Like what do you have fond memories of the guy? Anything come to mind? Mm hmm. [/INST] A friend told me one time I could draw a computer on the whiteboard faster than anybody had ever met and I said, you should meet my dad. Like when I was a kid he'd come home and say I was driving by this bridge and I was thinking about it and he put a little piece of paper and you draw the whole bridge. He was a mechanical engineer and he would just draw the whole thing and then he would tell me about it and tell me how you would have changed it. And he had this, you know, idea that he could understand and conceive anything. And I just grew up with that. So that was natural. So if you know, like when I interview people, I asked him to draw a picture of something they did on the whiteboard and it's really interesting. Like some people draw a little box, you know, and then they'll say, and this talks to this and be like, that's just frustrating. And I had this other guy come in one time. He says, Well, I designed a floating point in this chip, but I'd really like to tell you how the whole thing works and then tell you how the floating point works inside of. Do you mind if I do that? He covered two white boards in like 30 minutes, hired </s><s>[INST] him. Like he was great. This craftsman. I mean, that's the craftsmanship to that. [/INST] Yeah, But also the mental agility to understand the whole thing, put the pieces in contacts, you know, real view of the balance of the design worked because if you don't understand it properly, when you start to draw it, you'll, you'll fill up half the white board with like a little piece of it and you know, like your ability to lay it out and understandable way. It takes a lot of understanding. So </s><s>[INST] and be able to zoom into the detail and then Zuma [/INST] to make really fast. </s><s>[INST] What about the impossible things that your dad believe that you can do anything. That's a weird feature for a craftsman. Yeah, it seems that that uh echoes in your own behavior. Like that's that's the, [/INST] well, it's not that anybody can do anything right now, right? It's that if you work at it, you can get better at it and there might not be a limit. And they did funny things like, like he always wanted to play piano. So, at the end of his life, he started playing the piano when he had Parkinson's. He was terrible. But he thought if he really worked out in this life may be the next life, you'd be better at it, </s><s>[INST] He might be onto something. Yeah. [/INST] People enjoy doing it. Yeah. It's pretty funny. </s><s>[INST] Do you think the Perfect is the enemy of the good and hardware and software engineering? It's like we were talking about Javascript a little bit and the messiness of the 10 day building process. [/INST] Yes. Let's, you know, creative tension, right? Mhm. So, creative tension is you have two different ideas that you can't do both, right? And and but the fact that you want to do both causes you to go try to solve that problem. That's the creative part. So, we're building computers. Like some people say we have the schedule and anything that doesn't fit in the schedule we can't do, right? So, they throw out the perfect Because they have a schedule. I hate that. All right. Then there's other people who say we need to get this perfectly right? And no matter what? More people more money, right? And there's a really clear idea about what you want and some people really articulating it. So, let's call that the Perfect. Yeah. All right. But that's also terrible because they never ship anything. You never had any goals. So now you have now you have your framework, you can't throw out stuff because you can't get it done today because maybe you get it done tomorrow or the next project, right? You can't. So you have to, I work with a guy that really like working with, but he over filtered his ideas </s><s>[INST] over filters. [/INST] He started thinking about something as soon as he figured out was wrong with it, he'd throw it out and then I start thinking about it and you know, you come up with an idea and then you find out what's wrong with it and then you give it a little time to set because sometimes, you know, you figure out how to tweak it or maybe that idea help some other idea. So idea generation is really funny. So you have to give your idea space like spaciousness of mind is key, but you also have to execute programs and get shift done. And then it turns out computer engineering is fun because it takes, you know, 100 people to build a computer, 200 or 300 whatever the number is and people are so variable about, you know, temperament and you know, skill sets and stuff that in a, in a big organization you find the people who love the perfect ideas and the people that want to get stuffed on yesterday and people like that come up with ideas and people like the let's say shoot down ideas and it takes the whole it takes a large group of </s><s>[INST] people, some are good at generating ideas, some are good at filtering ideas and then all in that giant mess you somehow, I guess the goal is for that giant mess of people to uh find the perfect path through the tension, the creative tension. But like how do you know when you said there's some people good at articulating what perfect looks like what a good design is. Like if you're sitting in a room and you have a set of ideas about like how to design a better processor, how do you know this is this is something special here, this is a good idea, let's try this. [/INST] Have you ever brainstormed idea with a couple of people that were really smart and you kind of go into it and you don't quite understand it and you're working on it and then you start, you know, talking about it, put it on the whiteboard, maybe it takes days or weeks and then your brain starts to kind of synchronize, it's really weird, like you start to see what each other's thinking and it starts to work like you can see it work like my talent and computer design is I can I can see how computers work in my head like really well and I know other people can do that too and when you're working with people that can do that, like it is kind of amazing experience. And then, and every once a while you get to that place and then you find the flaw, it was just kind of funny because you can, you can fool yourself, but the two of you kind of drifted along the direction that was useless. That happens to like you have to because you know, the nice thing about computers and there's always reduction practice. Like you come up with your good ideas and I've noticed some architects who really love ideas and then they work on them and they put it on the shelf, then go work on the next idea and put on the shelf, they never reduced to practice. So they find out what's good and bad because mom was every time I've done something really new by the time it's done, like, the good parts are good, but I know all the flaws, like </s><s>[INST] yeah, would you say your career just your own experiences, your career defined by mostly by flaws or by successes? Like [/INST] again, there's great tension between those. If you haven't tried hard, right, and done something new. Okay, all right, then you're, you're not going to be facing the challenges when you build it, then you find out all the problems with it </s><s>[INST] and but when you look back, you see problems, [/INST] okay. When I look back, um what do you think earlier in my career, like even if I was the second Alpha Chip, uh I was so embarrassed about the mistakes, I could barely talk about it and it was in the Guinness Book of World Records and it was the fastest processor on the planet. So it was, and at some point I realized that was really bad mental framework to deal with doing something new. We did a bunch of new things and some worked out great and some were bad and we learned a lot from it and The next one, we learned a lot that EV six also had some really cool things in it. I think the proportion of good stuff went up, but it had a couple of fatal flaws in it that were painful. And then </s><s>[INST] you learn to channel the pain into like pride, [/INST] not pride, really just realization about how the world works, how that kind of ideas that works. </s><s>[INST] Life is suffering, that's the reality. What [/INST] it's not. You know, the buddhist said that a couple other people are stuck on it. No, it's, you know, there's just kind of weird combination of good and bad, you know, light and darkness that you have to tolerate and you know, deal with. Yeah, there's definitely lots of suffering </s><s>[INST] in the world, depends on the perspective. It seems like there's way more darkness, but that makes the light part really nice. What uh, computing hardware or just any kind of even software design are you, do you find beautiful from your own work, from uh other people's work? That you're just we were just talking about the battleground of flaws and mistakes and errors but things that were just beautifully done. Is there something that pops to mind? [/INST] Well when things are beautifully done? Usually there's a well set thought outside of abstraction layers </s><s>[INST] because the whole thing works in unison nicely. [/INST] And when I say abstraction layer that means two different components. When they work together they work independently, they don't have to know what the other one is doing. </s><s>[INST] So that [/INST] yeah, so the famous one was the network stack, like there's a seven layer networks, you know, data transport and protocol and all the layers and the innovation was is when they really wrote got that right because networks before that didn't define those very well. The layers could innovate independently and occasionally the layer boundary would you know, the interface would be upgraded and that that let you know the design space breathe, you could do something new and layer seven without having to worry about how they are four worked and so good design does that. And you see the processor designs when we did the zen design at AMG we made several components very modular and you know, my insistence at the top was I wanted all the interfaces defined before we wrote the article for the pieces. One of the verification leads had if we do this right, I can test the pieces so well independently when we put it together, we won't find all these interaction bugs because the floating point knows how the cash works. And I was a little skeptical, but he was mostly right that the the modularity designed greatly improved the quality. </s><s>[INST] Is that universally true in general, would just say about good designs. The modularity is like [/INST] usually talked about this before humans are only so smart. Like we're not getting any smarter right, But the complexity of things is going up. So, you know, a beautiful design can't be bigger than the person doing it. It's just, you know, their piece of it. Like the odds of you doing a really beautiful design of something that's way too hard for you is low, right? If it's way too simple for you, it's not that interesting. It's like, well anybody could do that. But when you get the right match of your, your expertise and you know, mental power to the right design size, that's cool. But that's not big enough to make a meaningful impact in the world. So now you have to have some framework to design the pieces so that the whole thing is big and harmonious. But you know, when you put it together, it's, you know, sufficiently sufficiently interesting to be used and you know, so that's like a beautiful design is </s><s>[INST] matching the limits of that human cognitive capacity to uh, to the module you can create and creating a nice interface between those modules and thereby do you think there's a limit to the kind of beautiful complex systems we can build with this kind of modular design is like uh you know, if we build increasingly more complicated you can think of like the internet. Okay, let's scale that you can think of like social networks like Twitter as one computing system. And but those are little modules, Right? So [/INST] it's built on it's built on so many components. Nobody at twitter even understands. So if an alien showed up and looked at twitter, he wouldn't just see twitter as a beautiful simple thing that everybody uses, which is really big. You would see the networks, it runs on the fiber optics. The data is transported the computers. The whole thing is so bloody complicated. Nobody twitter understands it. </s><s>[INST] So that's what the alien was. So yeah, if an alien showed up and looked at twitter or looked at the various different network systems that you can see on earth. [/INST] So imagine they were really smart, they could comprehend the whole thing. And then they sort of evaluated the human and thought this is really interesting. No human on this planet comprehends the system they built, </s><s>[INST] no individual. Well they even see individual humans. Like we humans are very human centric entity centric. And so we think of us as the organ as the central organism and the networks as just the connection of organisms. But from a perspective of and you know from an outside perspective, it seems [/INST] like when I get where the answer to the ant </s><s>[INST] colony, the ant colony? Yeah. Or the result of production of the ant colony, which is like cities and Yeah, it's it's uh in that sense, humans are pretty impressive the modularity that we're able to and the and how robust were to noise and mutation, all that kind of stuff because it's stress tested [/INST] all the time. Yeah. You know, you build all these cities with buildings and you get earthquakes occasionally and war some, you know, wars, earthquakes, </s><s>[INST] viruses every once in a while, [/INST] you know, changes in business plans for, you know, like shipping or something. Like, like as long as there's all stress test and then it keeps adapting to the situation. So, the yes, it's a curious phenomenon. </s><s>[INST] Well, let's go, let's talk about moore's Law a little bit uh at the broad view of moore's law was just exponential improvement of uh computing capability. Like open ai for example, recently published this kind of paper is looking at the exponential improvement in the training efficiency of neural networks for like, imagine that and all that kind of stuff. We just got better on this is purely software side just figuring out better tricks and algorithms for training neural networks and that seems to be improving uh significantly faster than the moore's law prediction? You know? So that's in the software space? What do you think if moore's law continues or if the general version of moore's law continues, Do you think that comes mostly from the hardware, from software? Some mix of the two. Some interesting totally uh not not the reduction of the size of the transistor kind of thing but more in the uh in the totally interesting kinds of innovations in the hardware space, all that kind of stuff. [/INST] Well there's like half a dozen things going on in that graph. So one is there's initial innovations that had a lot of room to be exploited. So you know the efficiency of the networks has improved dramatically and then the decompose ability of those in the U. S. You know, they started running on one computer that multiple computers and multiple Gps and raise the gps and they're up to thousands. And at some point. So so it's sort of like they were consumed, they were going from like a single computer application to 1000 computer application. So that's not really a moore's law thing. That's an independent vector. How many computers can I put on this problem? Because the computers themselves are getting better on like a moore's law rate but their ability to go from 1 to 10 to 100 to 1000 you know, it was something and then multiplied by, you know the amount of computer took to resolve like Alex net to resonate the transformers. It's it's been quite, you know, steady improvements. But </s><s>[INST] those are like escorts, aren't they? That's exactly kind of S curves that are underlying moore's law from the very [/INST] beginning. So </s><s>[INST] so what's the biggest, what's the most uh productive a rich source of S curves in in the future do you think is its hardware is and software. [/INST] So hardware is going to move along relatively slowly like you know, double performance every two years. There's still </s><s>[INST] like how you call that slow. [/INST] It's the slow version. The snail's pace with moore's law. Maybe we should we should uh we should trademark that one. </s><s>[INST] Whereas [/INST] the scaling by a number of computers, you know, can go much faster, you know, I'm sure at some point google had uh you know their initial search engine was running on a laptop, you know like and at some point they really worked on scaling that and then they factored the the indexer from you know this piece and this piece and this piece and they spread the data and more and more things and you know they did a dozen innovations. But as they scaled up the number of computers on that, it kept breaking finding new bottle of excellent their software and their schedulers and made them rethink like it seems insane to do a schedule her across 1000 computers who schedule parts of it and then send the results of one computer. But if you want to schedule a million searches, that makes perfect sense. So so there's the scaling by just quantity is probably the richest thing. But then as you scale quantity like a network that was great on 100 computers may be completely the wrong one. You may pick a network that's 10 times slower On 10,000 computers per computer. But if you go from 110,100 times. So that's one of the things that happened when we did internet scaling this efficiency went down. Not up the future of computing is inefficiency, not efficiency </s><s>[INST] but scales inefficient. [/INST] It's scaling faster than inefficiency bite you. And as long as there's you know dollar value, they're like scaling costs lots of money. But google showed facebook showed everybody showed that the scale was worth the money. Was that </s><s>[INST] it was, it was worth it financially. Do you think is it possible that like basically the entirety of Earth will be like a computing surface Like this table will be doing computing, this hedgehog will be doing computing. Like everything really inefficient. Don't computing [/INST] will be a picture book called computer roni um computation. We turn everything into computing. Well most of the elements aren't very good for anything. Like you're not gonna make a computer out of iron, like you know silicon and carbon have like nice structures. You know, we'll see what you can do with the rest of </s><s>[INST] it that people [/INST] talk about, well maybe we can turn the Sun into computer but it's it's hydrogen and a little bit of helium. So </s><s>[INST] what I mean is more like actually just adding computers to everything. [/INST] Okay, you're converting all the mass of the universe into computer. No, </s><s>[INST] so not [/INST] using the ironic from the simulation point of view is like the simulator build mass to simulate. Like </s><s>[INST] Yeah, I mean, yeah, so I mean, ultimately this is all heading towards the simulation. [/INST] Yeah, well, I think I might have told you this story. Tesla, they were deciding so they want to measure the current coming out of the battery and they decided between putting the resistor in there And putting a computer with a sensor in there and the computer was faster than the computer I worked on in 1982 and we chose the computer because it was cheaper than the resistor. So, so sure this hedgehog, you know, it cost $13 and we're gonna put a, you know, an ai that's as smart as you in there for five bucks. It will have one. Yeah. You know, so computers will be, you know, be everywhere. </s><s>[INST] I was hoping it wouldn't be smarter than me because [/INST] well everything's gonna be smarter than you. </s><s>[INST] But you're saying it's inefficient. I thought it's better to have a lot of dumb. [/INST] Well moore's law will slowly compact that </s><s>[INST] stuff, even the dumb things will be smarter than us. [/INST] The dumb things are gonna be smart. Are they going to be smart enough to talk to something that's really smart, you know, it's like, Well just remember like a big computer chip. You know, it's like an inch by an inch and you know 40 microns. Sick. It doesn't take very much very many items to make a high power computer. And 10,000 of them can fit in a shoebox. But you know, you have the cooling and power problems. But you know, people are working on that </s><s>[INST] but they still can't write a compelling poetry, music or uh understand what love is or have a fear of mortality. So so we're still winning. [/INST] Neither can most of humanity. So </s><s>[INST] well they can write books about it. So uh but but speaking about this uh this walk along the path of innovation towards uh the dumb things being smarter than humans. You are now the CTO Of uh can store two as of two months ago. They build hardware for deep learning. How do you build scalable and efficient? Deep learning. This is such a fascinating space. [/INST] Yeah. So it's interesting. So um up until recently I thought there was two kinds of computers. There's cereal computers that run like C. Programs and then there's parallel computers. So the way I think about it is, you know, parallel computers. You have given parallelism. Like GPS are great because you have a million pixels and modern Gpu s run a program on every pixel they call the shader program, right? So or like finite element analysis. You you build something, you know, you make this into a little tiny chunks to give each chunk to a computer. So you're giving all these chunks of parallelism like that. But most C programs you write this linear narrative and you have to make a go fast to make it go fast. You predict all the branches, all the data fetches, you run that more parallel but that's found parallelism. Um ai is I'm still trying to decide how fundamentalist is. It's a given parallelism problem but the way people describe the neural networks and then how they write them in pytorch makes graphs. </s><s>[INST] Yeah, that might be fundamentally different than the GPU kind [/INST] of parallelism. Yeah, it might be because when you run the GPU program on all the pixels you're running, you know, it depends, you know, this group of pixels say it's background blue and it runs a really simple program. This pixel is you know, some patch of your face. So you have some really interesting shader program to give you an impression of translucency but the pixels themselves don't talk to each other. There's no graph. Right? So you do the image and then you do the next image and you the next image and you run eight million pixels. Eight million programs every time in modern Gpus have like 6000 thread engines in them. So you know, to get eight million pixels each one runs a program on 10 or 20 pixels and that's how that's how they work. There's no graph. </s><s>[INST] But you think graph might be a totally a new way to think about hardware. [/INST] So Roger Federer and I've been having this conversation about giving versus found parallelism and then the kind of walk as we got more transistors like, you know, computers way back. When did stuff on scale or data. Now we did on vector data. Famous vector machines. Now we're making computers that operate on matrices. Right. And then the category we said it was next was spatial. Like imagine you have so much data that you know, you want to do the computer on this data. And then when it's done it says send the results of this pile of data on some software on that. And it's better to to think about it. Especially then to move all the data to a central processor and do all the work. </s><s>[INST] Especially. I mean moving in the space of data as opposed to moving the data, [/INST] you know, you have a you have a petabyte data space spread across some huge array of computers. And when you do a computation somewhere you send the result of that computation or maybe a pointer to the next program to some other piece of data and do it. But I think a better word might be graph and all the ai neural networks or graphs do some computations and the result here to another computation, do a data transformation to emerging to a pooling to another computation. </s><s>[INST] Is it possible to compress and say how we make this thing efficient? This whole process efficient. There's different. [/INST] So first the fundamental elements of the graphs are things like matrix multiplies, convolutions, data manipulations and data movements. So gps emulate those things with their little singles, you know, basically running the single threaded program and then there's, you know, in vehicles of the work where they grew up, a bunch of programs that are similar together. So for efficiency and instruction use and then at a higher level you kind of, You take this graph and you say this part of the graph as the matrix multiply which runs on these 32 threads but the model at the bottom was built for running programs on pixels, not executing </s><s>[INST] graphs. Emulation awesome. So it's impossible to build something that natively runs graphs. [/INST] Yes. So that's what started. So where </s><s>[INST] are we on that? How like in the history of that effort are we in the early days? Yeah. [/INST] So tense Torrance started by a friend of mine, Hlabisa by Jack and I was his first investor. So I've been, you know, kind of following them and talking to him about it for years and and the follow and I was considering things to do. I decided, you know that we, we held a conference last year with a friend organized it and and we wanted to bring in thinkers and two of the people were. Andre. Capaci And Chris Latner and Andre gave this talk show on YouTube called Software 2.0 which I think is great. Which is we went from programs computers where you write programs to data program computers you know like the futures you know of software is data programs the networks and I think that's true. And then um chris has been worked he worked on L O V E M. The low level virtual machine which became the intermediate representation for all compilers. And now he's working on another project called M L. I. R. Which is mid level intermediate representation which is essentially under the graph about how do you represent that kind of computation and then coordinate large numbers of potentially heterogeneous computers. And I would say technically tense torrents. You know two pillars are those those those two ideas software 2.0 and mid level representation. But it's in service of executing graph programs. The hardware is designed to do </s><s>[INST] that so it's including the hardware piece. [/INST] And then the other cool thing is for a relatively small amount of money, they did a test chip and to production chips. So it's like a super effective teams. And and unlike some ai startups where if you don't build the hardware to run the software that they really want to do then you have to fix it by writing lots more software. So the hardware naturally does Matrix multiply com pollution the data manipulations and the data movement between processing elements that that you can see in the graph, which I think is all pretty clever and </s><s>[INST] that's what I'm working on now. So, uh, the, I think it's called the grace called processor introduced last year. It's uh, you know, there's a bunch of measures of performance, we're talking about horses, it seems to outperform 368 trillion operations per second. Seems to outperform in videos. Tesla T four system. So these are just numbers. What do they actually mean in real world perform like what are the metrics for you that you're chasing in your horse race? Like what do you care about? [/INST] Well, first, so the native language of, you know, people who write aI network programs is pytorch. Now pytorch tensorflow, there's a couple of others. </s><s>[INST] The pytorch is one over tensorflow is just, [/INST] I'm not an expert on that. I know many people have switched from tensorflow. Pytorch and there's technical reasons for it and </s><s>[INST] I use both both are still awesome. [/INST] Both are still awesome. </s><s>[INST] But the deepest love is for pytorch. Currently [/INST] there's more love for that and that that may change. So the first thing is when they write their programs, can the hardware execute it pretty much as it was written. Right? So pytorch turns into a graph, we have a graph compiler that makes that graph then it fractions the graph down. So you have big matrix multiply. We turned it into right sized chunks to run on the processing elements. It hooks all the graph up. It lays out all the data. There's a couple of mid level representations of it that are also simulate able so that if you are writing the code, you can see how it's gonna go through the machine, which is pretty cool. And then at the bottom is scheduled kernels, like math, data manipulation, data, movement kernels, which do this stuff. So we don't have to run, write a little program to do matrix multiply because we have a big matrix multiplier. Like there's no sympathy program for that, but that there is scheduling for that. Right? So the one of the goals is if you write a piece of pytorch code, that looks pretty reasonable, you should be able to compile it running on the hardware without having to tweak it and and do all kinds of crazy things to get performance. </s><s>[INST] There's not a lot of intermediate steps. It's running directly [/INST] As like on a GPU if you write a large matrix multiply naively, you'll get 5-10% of the peak performance of the GPU. Right? And then there's a bunch of there's a bunch of people published papers on this and I read them about what steps do you have to do and it goes from pretty reasonable while transpose one of the matrices. So you do road or not colum ordered, you know, block it so that you can put a block of the matrix on different sms, you know, groups of threads, but some of it gets in the little little details like you have to schedule. It's just so so you don't have registered conflicts. So the they call them coding ninjas, could I just love it to get to the optimal point. You either write a pre use a pre written library, which is a good strategy for some things or you have to be an expert in micro architecture to program it. </s><s>[INST] Right? So the position step is really more complicated with the [/INST] so our goal is if you write pytorch, that's good pytorch, you can do it now. There's as the networks are evolving, you know, they've changed from convolutional to make just multiply that. People are talking about conditional graphs. We're talking about very large matrices, they're talking about scarcity. We're talking about problems that scale across many, many chips. So the native, you know, data item as a pocket, like, so you send a packet to processor, it gets processed, it does a bunch of work and then it may send packets to other processors and they execute like a data flow graphs. Kind of methodology. Got it. We have a big network on chip and then 16, the next second ship has 16 Ethernet ports to hope lots of them together. And it's the same graph compartment across multiple chips. </s><s>[INST] So that's where the skill [/INST] comes in. So it's built to scale naturally. Now, my experience with scaling is as you scale, you run into lots of interesting problems. So scaling is a mountain to climb. So the hardware is built to do this and we're in the process of </s><s>[INST] is there software part to this with, with ethernet and all that? [/INST] Well, the, you know, the protocol at the bottom, you know, we sent, you know, it's an ethernet fi but the protocol basically says send a packet from here to there. It's all point to point the head of it says which processor descended to and we basically take a pocket off our on chip network, put anything that had it on it, send it to the other end, strip the header </s><s>[INST] off and send us the local [/INST] thing. It's pretty straightforward. </s><s>[INST] Human human interaction is pretty straight forward to, but we can get a million of us were just crazy stuff together. Can be fun. So is that the goal is scale. So like for example, I have been recently doing a bunch of robots at home for my own personal pleasure. Uh am I going to ever use 10 story or is this more for [/INST] there's all kinds of problems like there's small inference problems or small training problems or thanks for any problems. </s><s>[INST] What's the big goal? Is it the big training problems with the small training problems [/INST] or one of the goals is the scale from 100 million watts to two Megawatt, you know, so like really have some range on the problems and the same kind of AI programs work at all different levels. So that's all the natural, since the natural data item is a pocket that we can move around, it's built to scale. But so many people have, you know, small problems. Right. Right. But, but you know, </s><s>[INST] they're like inside that phone is a small problem to solve. So do you see test or potentially being inside a phone? [/INST] Well, the power efficiency of local memory, local computation in the way we built it is pretty good. And then there's a lot of efficiency on being able to do conditional graphs and Spar City. I think it's, it's for complicated networks. I want to go into small factor. It's quite good. But we have to prove that that's all, </s><s>[INST] it's a fun problem. And that's the early days of the company writes a couple of years you said. But you think universities, you think they legit as you join? Well, that's, [/INST] well, it's also, it's a really interesting place to be like I, I was exploding, you know, and I looked at some other opportunities like build a faster processor, which people want, but that's more on incremental path than what's going to happen in a I in the next 10 years. So this is kind of, you know, an exciting place to be part of </s><s>[INST] the revolution will be happening in the very space that [/INST] lots of people working on it, but there's lots of technical reasons why some of them, you know, aren't going to work out that well. And, and you know, that's, that's interesting. And there's also the same problem about getting the basics right. Like we've talked to customers about exciting features and at some point we realized that we should never realizing they want to hear first about memory, bandwidth, local band with compute intensity program ability. They want to know the basics, power management, how the network ports work. What are the basics do all the basics work? Because it's easy to say, we've got this great idea that, you know, the crack gpt three. Uh, but the people we talked to understand if I buy the, so we have a piece of express card with our chip on it. If you buy the card, you plug it in your machine to download the driver. How long does it take me to get my network to run? Right. Right. You know, that's the real </s><s>[INST] question is a very basic question. So yeah, is there an answer to that yet? Or is it [/INST] trying to goals like an hour? </s><s>[INST] Okay. When can I buy a test for it? [/INST] Pretty </s><s>[INST] soon for my, for the small case [/INST] training? Pretty soon months. </s><s>[INST] Good. I love the idea of you inside a room with Carpathia? The Carpathia and chris Lattner? Uh very um, very interesting, very brilliant people, very out of the box thinkers, but also like first principles thinkers [/INST] Well, they both get stuff done, they only get stuff done to get their own projects done. They talk about it clearly to educate large numbers of people and they've created platforms for other people to go do their stuff on. </s><s>[INST] Yeah, the clear thinking that's able to be communicated. It's kind of impressive. [/INST] It's kind of remarkable to Yeah, I'm a fan. </s><s>[INST] Well, let me ask, because I talked to Chris actually a lot these days, he's been uh one of the just to give him a shout out and he's been so supportive as a human being. So everybody's quite different. Like, great engineers are different, but he's been like, sensitive to the human element in a way that's been fascinating. Like, he was one of the early people on this stupid podcasts that I do to say, like, don't quit this thing, and also talk to whoever the hell you want to talk to, that kind of from a legit engineer to get like props and be like, you can do this. That was, I mean, that's what a good leader does, Right? So it's just kind of let a little kid do his thing, like, go, go do it, let's see, let's see, see what turns out that that's a that's a pretty powerful thing. But what do you uh what's your sense about? He used to be now, I think stepped away from google, right? He said Science five, I think uh what's really impressive to you about the things that Chris has worked on, because it's that we mentioned the optimization, the compiler design stuff that LLVM. Uh then there's he's also a google, worked at the tpu stuff. Uh he's obviously worked on Swift, so the programming language side talking about people that work in the entirety of the stack, uh what from your time interacting with chris and knowing the guy, what's really impressive to you, it just inspires you. [/INST] Well, well, like L O V and became, you know, the platform, the de facto platform for compilers, like it's amazing and, you know, it was good code quality, good design choices, He hit the right level of abstraction, there's a little bit of the right time, the right place. And then he built a new programming language called Swift, which, you know, after, you know, let's say some adoption resistance became very successful. I don't know that much about his work at google, although I know that, you know, that was a typical, they started tensorflow stuff and, you know, it was new, it was, you know, they wrote a lot of code and then at some point it needs to be re factored to be, you know, because its development slowed down why pytorch started a little later and then passed it. So he did a lot of work on that and then his idea about M L I R, which is what people started to realize that the complexity to software stack above the low level ir it was getting so high that forcing the features of that into the level was putting too much of a burden on it. So he's splitting that into multiple pieces. And that was one of the inspirations for our software stack where we have several intermediate representations that are all execute Herbal and you can look at them and do transformations on them before you lower the level. So that was I think we started before M. O. I. R. Really got, you know, far enough along to use but we're interested in that </s><s>[INST] he's really excited about malaria. He's that's just like a little baby. So he and there seems to be some profound ideas on that that are really useful. [/INST] So so each one of those things has been, as the world of software gets more and more complicated, how do we create the right abstraction levels to simplify it in a way that people can now work independently on different levels of it. So I would say all three of those projects, all of them Swift and ml ir did that successfully. So I'm interested was what he's going to do next in the same kind of way. </s><s>[INST] So on either the Tpu or maybe the NVIDIA Gpu side, How does 10 story you think or the ideas underlying it doesn't to be tested and just this kind of graph focused uh graph centric hardware, deploring center hardware beat and videos do you think it's possible for it to basically overtaken video? What's what's that process look like? What's that journey looked like you think? [/INST] Well Gps were built to run shader programs on millions of pixels not to run graphs. Yes. So there's a hypothesis that says the way the graphs are built is going to be really interesting to be inefficient on computing this. And then the primitives is not a simple program. It's matrix multiply convolution. And then the data manipulations are fairly extensive about like how do you do a fast transpose with a program? Have you ever written that transports program? They're ugly and slow. But in hardware you can do really well. Like giving example. So when Gpu accelerators, so I started doing triangles, like you have a triangle which maps on the set of pixels. So you build it's very easy straightforward to build a hardware engine that will find all those pixels. It's kind of weird because you walk along the triangle to to get to the edge and then you have to go back down to the next row and walk along and then you have to decide on the edge if the line of the triangle is like half on the pixel, what's the pixel color because it's half of this pixel and half the next one. That's called rationalization </s><s>[INST] because you're saying that can be done in uh in [/INST] hard that's an example of that operation as a software program is really bad. I've written a program that addressed or ization the hardware that does it. It's actually less code than the software program that does it. And it's way faster. Right? So there's certain times when the abstraction you have rast arise a triangle. You know, execute a graph, you know, components of a graph that the right thing to do in the hardware. Software boundary is for the hardware to naturally do it. </s><s>[INST] So, the GPU is really optimized for the restoration of triangles. [/INST] Well, no, that's just well, like in a modern, you know, that's a small piece of modern Gpu s. What they did is that they still restaurant strangles when you're running in the game. But for the most part, most of the computation in the area, the Gpu is running Shader programs but they're single threaded programs on pixels, not crafts. </s><s>[INST] Let's be honest. Let's say. I don't actually know the math behind shader shading and lighting and all that kind of stuff. I don't know what [/INST] They look like. Little simple floating point programs or complicated ones. You can have 8000 instructions in a shader program, </s><s>[INST] but I don't have a good intuition why it can be parallelized so easily [/INST] Now because you have eight million pixels in every single, so you have a light, right? Yeah. That comes down the angle, you know, the amount of light, like, like say this is a line of pixels across this table. Right? The amount of light on each pixel is subtly different </s><s>[INST] and each pixel is responsible for figuring out [/INST] what figure it out. So that pixel says I'm this pixel, I know the angle of the light, I know the inclusion of the color. I am like every single pixel here is a different color. Every single pixel gets a different amount of light. Every single pixel has a subtly different translucency. So to make it look realistic, the solution was you run </s><s>[INST] a separate program on every pixel but I thought there's a reflection from all over the place, every [/INST] picture but there is so you build a reflection map which also has some pixelated thing. And then when the pixels looking at the reflection map has to calculate what the normal off the surfaces and it does it per pixel by the way, there's boatloads of hacks on that you like, you may have a lower resolution light map, reflection map. There's all these tax they do. But at the end of the day it's per pixel computation </s><s>[INST] and it so happened that you can map graph like computation onto the this pixel central company could [/INST] do floating point programs on convolutions and the matrices and and and video invested for years in CUDA first for HPC and then they got lucky with the AI </s><s>[INST] trend. But do you think they're going to essentially not be able to hardcore pivot out of there? We'll see. [/INST] That's always interesting. How often do big companies, hardcore pivot occasionally? </s><s>[INST] How much do you know about in video folks? So yeah. Well I'm curious as well who's ultimately as uh, [/INST] well they've, they've innovated several times, but they've also worked really hard on mobile. They worked really hard on radios, you know, you know, they're fundamentally a GPU company. </s><s>[INST] Well, they tried to pivot is an interesting little uh, game and play in autonomous vehicles right? With or a semi autonomous, like playing with Tesla and so on. And seeing that's a dipping a toe into that kind of pivot. [/INST] They came out with this platform, which is interesting technically, but it was like a 3000 watt you know, 1000 $3000 GPU platform. </s><s>[INST] I don't know if it's interesting technical, it's interesting philosophically. I technically, I don't know if it's the execution that craftsmanship is there. I'm not sure, but I didn't think they [/INST] were repurposing GPS for an automotive solution. </s><s>[INST] It's not a real pivot. [/INST] They didn't, they didn't build a ground up solution right? Like the, like the chips inside Tesla are pretty cheap like Mobileye has been doing this. They're they're doing a classic work from the simplest thing, you know, they were building 40 million square millimeter chips and then video, their solution had to 800 millimeter chips and to 200 millimeter chips and you know, like polos are really expensive. The rams and and you know, it's a really different approach. The mobile I fit the, let's say automotive cost and form factor and then they added features as it was economically viable and videos had take the biggest thing and we're gonna go make it work, You know? And, and that's also influenced like Waymo. There's a whole bunch of autonomous startups where they have a 5000 watt server in your trunk. Right? And, but that's That's because they think, well 5000 watts and you know, $10,000 is okay because it's replacing the driver Elon's approach was that board has to be cheap enough to put it in every single test. So whether they turn on autonomous driving or not, which mobili was like, We need to fit in the bomb and cost structure that car companies do. So they may sell your gps for $1,500. But the bomb for that's like $25 </s><s>[INST] well. And uh, for mobile I it seems like neural networks were not first class citizens like the computation, they didn't start out as a, [/INST] uh, it was a seabee problem. Yeah. And the classic tv and found stoplights and lines and they were really good at it. </s><s>[INST] Yeah. And they never, I mean, I don't know what's happening now, but they never fully pivoted. I mean it's like it's the video thing then as opposed to, so if you look at the new Tesla work, it's like neural networks from the ground up. [/INST] Yeah. And even Tesla started with a lot of tv stuff in it. And Andre has basically been eliminated, you know, move and move everything into the network. </s><s>[INST] So without this isn't like confidential stuff, but you sitting on a porch looking over the world, looking at the work that Andre is doing that Ellen is doing with Tesla autopilot. Do you like the trajectory of where things are going on [/INST] there? Making serious progress? I like the videos that people are driving debate of stuff like it's taken some pretty complicated intersections and all that, but it's it's still an intervention for dr I mean I I have all of the current autopilot. My Tesla, I use it every day. </s><s>[INST] Do you have full self driving beta or not? So you like where this [/INST] is going? We're making progress. It's taking longer than anybody thought, you know, my my wonder was, is, you know, hardware three, is that enough computing off by two off by five by 10 off by 100. And and I thought it probably wasn't enough, but they're doing pretty well with it now. and one thing is the days that gets bigger, the training gets better. And then there's this interesting thing is you sort of train and build an arbitrary size network that solves the problem. And then you? Re factor the network down to the thing that you can afford the ship. Right? So the, the goal isn't to build a network that fits in the phone, it's to build something that actually works. Mhm. And then then how do you make that most effective on the hardware you have? And they seem to be doing that much better than a couple of years ago? </s><s>[INST] Well, the one really important thing is also what they're doing well is how to iterate that quickly, which means like it's not just about one time deployment. One building is constantly hearing the network and trying to automate as many steps as possible. Right? And that's actually the principles of the software to like you mentioned with Andre is uh It's not just, I mean, I don't know what the actual his description of software 2.0 is if it's just high level fill software, their specifics. But the interesting thing about what that actually looks in the real world is is that what I think Andre calls the data engine, It's like, it's the iterative improvement of the thing. You have a neural network that does stuff fails on a bunch of things and learns from it over and over and over. So you're constantly discovering edge cases. So it's very much about uh like data engineering, like figuring out it's kind of what you were talking about with 10 stories, you have the data landscape, They have to walk along that data landscape in a way that uh it's constantly improving the neural network and that that feels like that's the central pieces. So [/INST] there's two pieces of it, like you find education that don't work and then you define something that goes get your data for that. But then the other constraint is whether you have to label it or not, like the amazing thing about like the GPT three stuff is it's unsupervised, so there's essentially infinite amount of data now, there's obviously infinite amount of data available from cars of people successfully driving. But you know, the current pipelines are mostly running on labeled data, which is human limited. So when that becomes a new unsupervised, right, it will create unlimited amount of data, which is no scale. Now the networks that they use that data might be way too big for cars, but then it will be the transformation from now we have unlimited data. I know exactly what I want. Now, can I turn that into something that fits in the car And that that process is going to happen all over the place every time you get to the place we have unlimited data And that's what software 2.0, is about unlimited data training networks to do stuff without humans writing code to do it. </s><s>[INST] And ultimately also trying to discover, like you're saying this self supervised formulation of the problem. So the unsupervised formulation of the problem, like uh you know, in driving, there's this really interesting thing, which is you look at a scene that's before you and you have data about what a successful human driver did in that scene. You know, one second later, it's a little piece of data that you can use just like with GPT three as training currently, even though Tesla says they're using that, it's an open question to me how much, how far can you, can you saw all of the driving with just that self supervised piece of data and like I I think [/INST] that's what comedy, </s><s>[INST] that's what I'm a i is doing. But the question is how, how much data? So we're comedy doesn't have is as good of a data engine for example, as Tesla does. That's where the like the organization of the data. Uh, I mean as far as I know, I haven't talked to George but they do have the data. The question is how much data is needed because we say infinite very loosely here. It's And then the other question which you said, I don't know if you think it's still an open question is are we in the right order of magnitude for the compute necessary? Um, that is this is it like what Elon said this chip that's in there now is enough to do full self driving or doing another order of magnitude. I think nobody actually knows the answer to that question. I like the confidence that Ellen has. But [/INST] yeah, we'll see there's another funny thing is you don't learn to drive with infinite amounts of data. You learned to drive with an intellectual framework that understands physics and color and horizontal services and laws and roads and you know, all your, your experience from manipulating your environment. Like look, there's so many factors go into that. So and then when you learn to drive, driving is a subset of this conceptual framework that you have. Right? And so self driving cars right now, we're teaching them to drive with driving data. You never teach you human to do that. You teach human all kinds of interesting things like language like don't do that, you know, watch out. You know, there's all kinds of stuff going on. </s><s>[INST] This is where you I think previous time we talked about where you poetically uh disagreed with my naive notion about humans. I just think that humans will make this whole driving thing really difficult. [/INST] Yeah. All right. I said he was don't move that slow. The ballistics problem. </s><s>[INST] The ballistic humans are ballistics problem, which is like poetry to me, it's very it's very possible that in driving there. Indeed, purely a ballistics problem. I and I think that's probably the right way to think about it. But I still they still continue to surprise me. Those damn pedestrians, cyclists, other humans and other cars. [/INST] And yeah, it's gonna be one of these compensating things. So like when you're driving, you have an intuition about what humans are going to do, but you don't have 360 cameras and radars and you have an intention problems. So, so, so the self driving car comes in with no attention problems. 3 60 cameras, you know, a bunch of other features. So they'll wipe out a whole class of accidents. Right? And you know, you know, emergency braking with radar and especially as it gets, you know, ai enhanced will eliminate collisions. Right? But then you have the other problems of these unexpected things where, you know, you think you're human intuition is helping. But then the cars also have, you know, a set of hardware features that you're, you're not even close to. </s><s>[INST] And the key thing, of course, is, uh, if you wipe out a huge number of kind of accidents and it might be just way safer than a human driver. Even though even if humans are still a problem, that's hard to figure out. [/INST] Yeah, that's, that's probably what happened is autonomous cars will have a small number of accidents humans would have avoided. But the wipe, they'll get rid of the bulk of them. </s><s>[INST] What do you think about, uh, like Tesla's Dojo efforts or it can be bigger than Tesla in general. It's kind of like the tents torrent. Uh, trying to innovate? Like this is the, the economy like should accompany trying to from scratch, build its own neural network training [/INST] hardware. Well, first, I think it's great. So we need lots of experiments. Right? And there's lots of startups working on this and are pursuing different things. And I was there when we started dojo and it was sort of like what's the unconstrained computer solution to go do very large training problems and then there's fun stuff like, you know, he said, well we have this 10,000 watt bored too. Cool. Well you go talk to guys that SpaceX and they think 10,000 watts is a really small number, not a big number and and there's brilliant people working on it. I'm curious to see how it'll come out. I couldn't tell you, you know, I know it pivoted a few times since I left. So, </s><s>[INST] so the cool existing to be a big problem. I do like what, you know, and said about it, which is like we don't want to do the thing unless it's way better than the alternative, whatever the alternative is. So it has to be way better than like racks of GPS. [/INST] Yeah. And then the other thing is just like, you know, you know, the Tesla autonomous driving hardware, It was only serving one software stack and the hard work came in, the software team were tightly coupled. You know, if you're building a general purpose ai solution and you know, there's so many different customers with so many different needs Now, something Andre said is I think this is amazing 10 years ago like vision recommendation language were completely different disciplines. He said the people literally couldn't talk to each other. And three years ago, it was all neural networks but the very different neural networks and recently it's converging on one side. The networks, they vary a lot in size obviously the variant data variant outputs but the technology has converged a good bit. </s><s>[INST] Yeah, these transformers behind GPT three, it seems like they could be applied to video that could be applied to a lot of it's like and they're all really, [/INST] it was like to literally replace letters with pixels, it does vision, it's amazing. </s><s>[INST] So and then size actually improves the thing so the bigger it gets, the more completely throughout it, the better it gets [/INST] and the more data you have, the better it gets. So so then you start to wonder, well is that a fundamental thing or is this just another step to some fundamental understanding about this kind of computation which is really interesting </s><s>[INST] as humans don't want to believe that that kind of thing will achieve conceptual understanding as you were saying like you'll figure out physics but maybe it will, maybe probably will. [/INST] Well it's worse than I thought it will understand physics in ways that we can't understand. I like your Stephen Wolfram talk where he said, you know, there's three generations of physics, there was physics by reasoning Well, big things should false faster than small things. Right, that's reasoning and then there is physics by equations like you know, but the number of programs in the world that are solved with a single equation is relatively low. Almost all programs have, you know, more than a one line of code, maybe 100 million lines of code. So he said that now we're going to physics by equation which is his project, which is cool. I might point out that there was there was 22 generations of physics before a reasoning habit. Like all animals, you know, know, things fall and you know, birds fly and you know, predators know how to, you know, solve a differential equation to cut off, accelerating, you know, curving animal path. And then there was, you know, the gods did it. Right, So right. So there, you know, there's five generations now. Software 2.0 says programming things is not the last step data. So there's going to be a physics power stevens. Oh, friends come, </s><s>[INST] that's not explainable, explainable. [/INST] And and actually there's no reason that I can see while that even that's the limit. Like there's something beyond that. I mean they're like usually when you have this hierarchy, it's not like, well, if you have this step and this step and this step and they're all qualitatively different and conceptually different. It's not obvious why, you know, six is the right and number of hierarchy steps and not seven or eight or </s><s>[INST] well, then it's probably impossible for us to to comprehend something that's beyond the thing that's not explainable. [/INST] Yeah, but I think, but the thing that, you know, understands the thing that's not explainable to us well conceived the next one and like I'm not sure why there's a limit to it. </s><s>[INST] Uh [/INST] click your brain hurts. It's </s><s>[INST] just hard story. If we look at our own brain which is an interesting uh illustrative example in your work with Henthorn and try and design deep learning architectures. Uh Do you do you think about the brain at all maybe from hardware designer perspective, if you could change something about the brain, what would you change? Or do a funny question, [/INST] How would you know, your brain is really weird? Like, you know your true real cortex where we think we do most of our thinking Is what like six or 7 neurons thick. Like that's weird. Like all the big networks are way bigger than that, like way deeper. So that seems odd. And then, you know, when you're thinking, if it's if if the input generates a result, you can lose, it goes really fast. But if you can't that generates an output, that's interesting. Which turns into an input and then your brain to the point where you mull things over for days and how many trips through your brain is that? Right? Like it's, you know, 300 milliseconds or something to get through seven levels of neurons. I forget the number exactly. But then it does it over and over and over as it searches and the brain clearly it looks like some kind of graph because you have a neuron with you know connections and it talks to other ones and it's locally very computational intense. But it's also does sparse computations across a pretty big area. </s><s>[INST] There's a lot of messy biological type of things. And it's it's meaning like first of all there's mechanical, chemical and electrical signals, that's all that's going on. Then the there's uh a synchronicity of signals and there's like there's just a lot of variability that seems continuous and messy and just a mess of biology and it's unclear whether that's a good thing or it's a bad thing because if it's a good thing that we need to run the entirety of the evolution, well we're gonna have to start with the basic bacteria to create [/INST] Something. We could you could build a brain with 10 layers. Would that be better or worse or more more connections or less connections or you know we don't know to what level our brains are optimized? But if I was changing things like like you know you can only hold like seven numbers in your head like why not 100 or a million? </s><s>[INST] Is [/INST] that why can't like why can't we have like a floating point processor that can compute anything we want like and see it all properly. Like that would be kind of fun. And why can't we see in four or 8 dimensions. Like like three D. is kind of a drag like all the hard mass transforms are up in multiple dimensions. So there's that, you know, you could imagine a rain architecture that you know, you could enhance with a whole bunch of features that would be you know, really useful for thinking about things. </s><s>[INST] It's possible that the limitations you're describing are actually essential for like the constraints are essential for creating like the depth of intelligence like that. The ability to reason. [/INST] Yeah, it's hard to say because Like your brain is clearly a parallel processor. You know, you know, 10 billion neurons talking to each other at a relatively low clock, right? But it produces something that looks like a serial thought process. It's a serial narrative in your head. That's true. But then there are people famously who are visual thinkers. Like I think I'm a relatively visual thinker. I can imagine any object and rotated my head and look at it. And there are people who say they don't think that way at all. And recently I read an article about people, people who say they don't have uh they don't have a voice in their head, they can talk. But when they, you know, it's like, well what are you thinking? They'll describe something that's visual. So that's curious now if you're saying If we dedicated more hardware to holding information like you know, 10 numbers or a million numbers. Like would that just distract us from our ability to form this kind of </s><s>[INST] singular identity dissipates somehow. [/INST] Right? But but maybe you know future humans will have many identities that have some higher level organization but can actually do lots more things in parallel. </s><s>[INST] Yeah. There's no reason if we're thinking module early, there's no reason we can't have multiple consciousnesses in one brain [/INST] and maybe there's some way to make it faster so that the you know the the area the computation could could still have unified feel to it while still having way more ability to do parallel stuff at the same time. Could definitely be improved. </s><s>[INST] Could be improved. Well it's pretty good right now actually people don't give it enough credit. The thing is pretty nice. You know the fact that the right end seemed to be and give a nice like spark of ah beauty to the whole experience. I don't know. I don't know if it can be improved easily. [/INST] It may be more beautiful. </s><s>[INST] I don't know how. Yeah. What [/INST] do you mean? What do you mean? How All the ways you can't imagine. </s><s>[INST] No, but that's the whole point I wouldn't be able to um at the fact that I can imagine ways in which it could be more beautiful. [/INST] So you know you know Ian banks his stories so the super smart ai is there live mostly live in the world of what they call infinite fun because they can create arbitrary world. So they interact in you know the story has it? They interact in the normal world and they're very smart and they can do all kinds of stuff and you know given mind can, you know, talk to a million humans at the same time as we're very slow and for reasons artificial the story, they're interested in people and doing stuff but they mostly live in this, this other land of thinking. </s><s>[INST] My inclination is to think that the ability to create infinite fund will will not be so fun. [/INST] That's sad. There's so many things to do. Magic. Be able to make a star planets around. </s><s>[INST] Yeah, Yeah. But because we can imagine that is why life is fun. If we can, if we actually were able to do it, it'll be a slippery slope. We're fun would even have a meeting because we just consistently desensitize ourselves by the infinite amounts of fun we're having and the sadness, the dark stuff is what makes it fun. I think that could be the Russian, [/INST] it could be the, could be the fun, makes it fun and sadnesses makes it bittersweet. </s><s>[INST] Yeah, that's true fun. Could be the thing that makes it fine. So what do you think about the expansion? Not through the biology side, but through the Bc I the brain computer interfaces. Yeah, you got a chance to check out the neuralink stuff. [/INST] It's super interesting. Like, like humans like like our thoughts to manifest this action, you know, like, like as a kid, you know like shooting a rifle was super fun, driving a mini bike, doing things and then computer games, I think for a lot of kids became the thing where they, you know, they can do what they want, they can fly a plane, they can do this, they can do this, right? But you have to have this physical interaction now, imagine, you know, you could just imagine stuff and it happens right? Like really richly and interestingly like we kind of do that when we dream. Like dream dreams are funny because like if you have some control or awareness in your dreams like it's very realistic looking or not realistic depends on the dream, but you can also manipulate that and you know what's possible? There is it is odd and the fact that nobody understands it's hilarious. But </s><s>[INST] um do you think it's possible to expand that capability through computing? Sure. Is there some interesting. So from a hardware designer perspective is there do you think will present totally new challenges and the kind of hardware required that like so this hardware isn't stand alone computing. [/INST] Just networking with the brain today, computer games are rendered by Gps. Right. Right. So, but you've seen the gang stuff right? Where train neural networks render realistic images, but there's no pixels, no triangles, no shade ear's no light maps, No nothing. So the future of graphics is probably a i right now that AI is heavily trained by lots of real data. Right? So if you have an interface with a Ai renderers, right? So if you say render a cat, it won't say well, how tall is the cat and how big it will render a cat? And you might say a little bigger, a little smaller, you know, make it a tabby shorter hair. You know, like you could tweak it like the amount of data you have to it sent to interact with a very powerful ai render could be low. </s><s>[INST] But the question is brain, computer interfaces would need to render not onto a screen but render onto the brain and like directly. So there's a [/INST] band do it both ways. I mean our eyes are really good sensors, it could render onto a screen and we could feel like we're participating in it. You know, they're going to they're going to have you know, like the Oculus kind of stuff. It's going to be so good when projections in your eyes you think it's real, you know, they're slowly solving those problems. And I suspect when the renderers of that information into your head is also a I mediated, they'll be able to give you the cues that you know, you really want for depth and all kinds of stuff. Like your your your brain is partly faking your your visual field right? Like your eyes are twitching around but you don't notice that occasionally they blank. You don't notice that, you know, there's all kinds of things like you think you see over here, but you don't really see there. It's all fabricated. So </s><s>[INST] peripheral vision is fascinating. [/INST] So if you have an ai render that's trained to understand exactly how you see and the kind of things that enhance the realism of experience could be super real actually. So I don't know what the limits that are. But obviously if we have a brain interface that goes in inside your visual cortex in a better way than your eyes do, which is possible for a lot of neurons. Yeah, maybe that will be even cooler. </s><s>[INST] But the really cool thing is it has to do with the infinite fund that you're referring to, which is our brain seems to be very limited. And like you said, computation is [/INST] very plastic. Very plastic. And so it's a it's a interesting combination. </s><s>[INST] The the interesting open question is the limits of that in your plasticity. Like how how flexible is that thing? Because we don't we haven't really tested it. [/INST] We know about that experiments where they put like a pressure pad on somebody's head and had a visual transducer pressurize it. And somebody slowly learned to see. It's like it's </s><s>[INST] especially at a young age if you throw a lot at it. Like what what can it uh can it completely? So can you like arbitrarily expanded with computing power. So connected to the internet directly [/INST] somehow. Yes, probably. Yes. </s><s>[INST] So the problem with Biology and ethics is like there's a mess. They're like us. Humans are perhaps unwilling to take risks in uh, in two directions that are full of uncertainty. [/INST] So 90% of the population is unwilling to take risks. The other 10% is rushing into the risks unaided by any infrastructure whatsoever. And you know, that that's where all the fun happens in society has been huge transformations in the last, you know, a couple of 1000 years. </s><s>[INST] Yeah, it's funny. I mean, I got a chance to interact with this Matthew johnson from johns Hopkins, he's doing this large scale study of psychedelics. It's becoming more and more. I've gotten a chance to interact with that community of scientists working on psychedelics, but because of that, that opened the door to me to all these uh, what do they call psycho notes, the people who, like you said, the 10% who like, I don't care. I don't know if there's a science behind this. I'm taking this spaceship. Uh, if I'm being the first on mars, I'll be uh, you know, you know, psychedelics interesting in the sense that in another dimension, uh, like you said, it's a way to explore the limits of the human mind. Like what is this thing capable of doing? Because you kind of like when you dream, you detach it. I don't know exactly in your science of it, but you detach your like reality from what your mind, the images, your mind is able to conjure up in your mind goes into weird places. Like entities appear somehow Freudian type of like trauma is probably connected in there somehow. But you start to have like these weird vivid worlds that like, [/INST] so do you actively dream you? Why not? I have like 66 hours of dreams and I it's like a really useful time. </s><s>[INST] I know I I haven't, I don't for some reason I just knock out and uh I have sometimes like anxiety inducing kind of like very pragmatic like nightmare type of dreams but not nothing fun, nothing, nothing fun, nothing fun. I I try I unfortunately have mostly have fun in the waking world which is very limited in the amount of fun you can have. [/INST] It's not that limited either. Yeah, that's what we'll have to talk uh </s><s>[INST] new instructions. Uh [/INST] Yeah, there's like a manual for that. You might want to. </s><s>[INST] I looked it up, I'll ask you on what would you dream [/INST] you know, years ago and I read about, you know like you know a book about how to have, you know become aware of your dreams. I worked on it for a while. Like there's this trick about, you know, imagine you can see your hands and look out and and I got somewhat good at it like but my mostly when I'm thinking about things are working on problems I I prepped myself before I go to sleep. It's like I pull into my find all the things I want to work on or think about and then that let's say greatly improves the chances that I'll work on that while I'm sleeping. And then, and then I also, you know, basically asked to remember it and I often remember very detailed </s><s>[INST] within the dream or outside the [/INST] dream well to bring it up in my dreaming and then remember when I wake up, it's just, it's more of a meditative practice you say, you know, to prepare yourself to do that. Like if you go to the sleep, still gnashing your teeth about some random things that happened that you're not that really interested in, you'll dream about it. </s><s>[INST] That's really interesting maybe. [/INST] But you can direct your dreams somewhat by prepping, </s><s>[INST] you know, I'm gonna have to try that. It's really interesting. Like the most important, the interesting, not like uh what did this guy send an email kind of like stupid worry stuff but like fundamental problems. You're actually concerned [/INST] about interesting things you're worried about books you're reading or you know, some great conversation you had or some some adventure you want to have. Like there's there's a lot of space there and and it seems to work that, you know, my percentage of interesting dreams and memories went up. </s><s>[INST] Is there? Uh is that the source of uh if you were able to deconstruct like where some of your best ideas came from, Is there a process that's at the core of that? Like so some people you know walk and think some people like in the shower the best ideas hit him if you talk about like newton apple hitting him on the head [/INST] now I I found out a long time ago I'm I processing so much slowly. So like in college I had friends who could study at the last minute get in the next day, I can't do that at all. So I always front loaded all the work like I do all the problems early you know for finals like the last three days I wouldn't look at a book because I want you know because like a new fact day before finals may screw up my understanding of what I thought I knew. So my my my goal was to always get it in and give it time to silk and I used to you know I remember we were doing like three D. Calculus, I would have these amazing dreams of three D surfaces for normal, you know calculating the gradient and just like all come up so it was really fun like very visual and uh and if I got cycles of that that was useful. Um And the other is don't over filter your ideas like I like that process of brainstorming where lots of ideas going to happen. I like people who have lots of ideas but then then yeah I'll let them sit and let it breathe a little bit and then reduce it to practice. Like at some point you really have to does it really work like it was this real or not right? But but you have to do both. There's great attention. They're like how do you be both open and you know, precise. </s><s>[INST] Have you had ideas that you just that sit in your mind for like years before the Sure it's an interesting way to he's generated ideas and just let them sit, let them sit there for a while. Mhm. I think I have a few of those ideas. [/INST] You know, it was so funny. Yeah, I think that's you know, creativity discipline or something </s><s>[INST] for the slow thinkers in the room I suppose because I some people like you said are just like like the [/INST] it's really interesting. There's so much diversity in how people think, you know, how fast or slow they are, how well they remember it. Don't look I'm not super good at remembering facts but processes and methods. Like in our engineer, I went to penn state and almost all our engineering tests were open book. I could remember the page and not the formula but soon as I saw the formula could remember the whole method if I if I learned it. Yeah. So it's just a funny Where some people could, you know, I had switched friends like flipping through the book trying to find the formula. Even knowing that they had done just as much work and I would just open the book, I was on page 27 about half, I can see the whole thing visually. </s><s>[INST] Yeah. You know, you have to learn that about yourself and figure out what the would function optimally. [/INST] I had a friend who was always concerned, he didn't know how he came up with ideas, had lots of ideas, but he said they just sort of popped up like you'd be working on something, I have this idea like where does it come from? But you can have more awareness of it, like, like, like, like how, how your brain works as a little murky as you go down from the voice in your head or the obvious visualizations. But when you visualize something, how does that happen? Yeah. You know if I say, you know, visualize volcano, it's easy to do, right? </s><s>[INST] And what does it actually look like when you visualize it? [/INST] I can visualize to the point where I don't see that very much out of my eyes and I see the colors of the thing I'm visualizing. </s><s>[INST] Yeah, but there's like a there's a shape, there's a texture, there's a color but there's also conceptual visualization. Like what do you actually visualizing when you're visualizing volcano? Just like peripheral vision? You think you see the whole [/INST] thing? Yeah, that's a good way to say it, you have this kind of almost peripheral vision of your visualizations, they're like these ghosts. But you know, if you if you work on it, you can get a pretty high level of detail </s><s>[INST] and somehow you can walk along those visualizations to come up with an idea which is uh [/INST] but weird, but when you're thinking about solving problems, like you're putting information and you're exercising the stuff you do know, you're sort of teasing the area that you don't understand and don't know, but you can almost, you know, feel, you know that process happening. You know? That's that's how I like, like like I know sometimes when I'm working really hard on something like like I get really hot when I'm sleeping and you know it's like we got the blanket throw, I wake up all the blankets on the floor and you know every time it's well we wake up and think, wow that was great, you know, are you able to </s><s>[INST] uh to reverse engineer what the hell happened there? [/INST] Sometimes it's vivid dreams and sometimes it's just kind of like you say like shadow thinking that you you sort of have this feeling you're you're going through this stuff but it's it's not that obvious </s><s>[INST] is that so amazing that the mind just does all these little experiments. I never you know, I thought I always thought it's like a river that you can't you're just there for the ride. But you're right if you prep it [/INST] it's all understandable meditation really helps. You got to start figuring out, you need to learn language of your own mind and there's multiple levels of it, </s><s>[INST] but the abstractions again, right? [/INST] It's somewhat comprehensible and observable and feel able or whatever the right word is. No, it's yeah, you're not along for the ride, You are the right. </s><s>[INST] I have to ask you hardware engineer working on your own networks now, what's consciousness? What the hell is that thing? Is that Is that just some little weird quirk of our particular computing device or is it something fundamental that we really need to crack open for two to build good computers? Do you ever think about consciousness? Like why it feels like something [/INST] to be? It's really weird. So yeah, I mean everything about it's weird. First, it's a half a second behind reality, right? It's a post hoc narrative about what happened, You've already done stuff by the time you're conscious of it And your consciousness generally is the single thread of the thing. But we know your brain is 10 billion neurons running some crazy parallel thing and there's a really big sorting thing going on there. It also seems to be really reflective in the sense that yeah, you create a space in your head, like we don't really see anything right? Like photons hit your eyes, it gets turned into signals that goes through multiple layers of neurons, you know, like I'm so curious that, you know, that looks glassy and that looks not glassy, like like how the resolution of your vision is so high. You have to go through all this processing where for most of it looks nothing like vision, like, like there's no theater in your mind, right? So we we have a world in our heads. We're literally just isolated behind our sensors. But we can look at it. Speculate about it, speculated about alternatives. Problem, solve what if you know, there's so many things going on and that process is lagging reality </s><s>[INST] and it's single threaded even though the underlying thing is like massively parallel. [/INST] So it's it's so curious. So imagine you're building ai computer if you wanted to replicate humans. Well you have huge rays of neural networks and apparently only six or seven deep. What's Claris they only remember seven numbers. But I think we can upgrade that a lot. Right? And then somewhere in there you would train the network to create basically the world that you live in. </s><s>[INST] Right? So tell stories to itself about the world that is proceeding [/INST] well create to create the world, tell stories in the world and then have many dimensions of you know, like side jobs to it. Like we have an emotional structure. Like we have a biological structure and that seems hierarchical to like, like if you're hungry, dominates your thinking, if you're mad, it dominates your thinking. Like and we don't know if that's important to consciousness or not, but it certainly disrupt, you know, intrudes in the consciousness like Children, lots of structure to that and we like to dwell on the past. We like to think about the future. We like to imagine we'd like to fantasize, right? And the somewhat circular observation of that is the thing we we call consciousness. Now, if you created the computer system and did all things create worldviews, created a future, alternate histories, you know, dwelled on past events accurately or semi accurately. You know, it's </s><s>[INST] it's consciousness [/INST] just bring up like, well with that feel look and feel conscious to you, like, you see just to me, but </s><s>[INST] external observers. Do you think a thing that looks conscious is conscious? Like, do you again, this is like an engineering kind of question, I think because uh, like, I don't know if we want to engineer consciousness. Is it okay to engineer something that just looks conscious or is there a difference between [/INST] we have all consciousness? Because it's a super effective way to manage our affairs </s><s>[INST] social? Well, [/INST] it gives us a planning system. You know, we have a huge amount of stuff like when we're talking like the reason we can talk really fast as we're modeling each other a really high level of detail </s><s>[INST] and consciousness is required for [/INST] that. Well, all those components together manifest consciousness. Right. So if we make intelligent beings that we want to interact with that were like, wondering what they're thinking, you know, looking forward to seeing them, you know, when they interact with them. They're interesting, surprising, you know, fascinating. You know, they will probably be feel conscious like we do and well perceive them as conscious. I don't know why not, but you never know </s><s>[INST] another fun question on this. Because in uh from a computing perspective, we're trying to create something that's human like or superhuman. Like let me ask about [/INST] aliens, aliens. </s><s>[INST] Uh do you think there's intelligent alien civilizations out there? And do you think their technology their computing their ai bots? There are their chips are of the same nature as ours? [/INST] Yeah, I got I have no idea. I mean, if there's lots of aliens out there that have been awfully quiet, you know, there's there's speculation about why there seems to be more than enough planet out there. </s><s>[INST] There's a [/INST] lot um there's intelligent life on this planet that seems quite different, you know, like, you know, dolphins seems like plausibly understandable octopuses don't seem understandable at all. If they live longer than a year, maybe they would be running the planet. They seemed really smart and there are neural architecture is completely different than ours. Who knows how they perceive things? </s><s>[INST] I mean, that's the question is for us intelligent beings, we might not be able to perceive other kinds of intelligence if they become sufficiently different than us. [/INST] We live in the current constrained world, You know, it's three dimensional geometry and the geometry defines a certain amount of physics and, you know there's like how time works seems to work. There's so many things that seemed like a whole bunch of the input parameters to the another conscious being are the same. Like it was biological biological things seem to be in a relatively narrow temperature range right? Because you know organics don't aren't stable too cold or too hot. So there's if you specify the list of things that input to that but soon as we make really smart beings and they go solve about how to think about a billion numbers at the same time and how to think and dimensions there's a funny science fiction book where the all the society had uploaded into this matrix and at some point on some of the beans and the Matrix thought I wonder if there is intelligent life out there. So they had to do a whole bunch of work to figure out like how to make a physical thing because they're Matrix was self sustaining and they made a little spaceship and they traveled to another planet when they got there there was like life running around but there was no intelligent life and then they figured out that there was this huge you know organic matrix all over the planet inside there were intelligent beings had uploaded themselves into that matrix. So everywhere intelligent life was soon as it got smart It up leveled itself into something way more interesting than three D. Geometry </s><s>[INST] and yeah escaped whatever this escape to blow is better. The essence of what we think of as an intelligent being. I tend to like the thought experiment of the organisms like humans aren't the organisms like the notion of like richard dawkins and memes that ideas themselves are the organisms like that are just using our minds to evolve. So like we're just like meat receptacles for ideas to breed and multiply and so on. And maybe those are the aliens. [/INST] So Jordan Peterson has line says, you know, you think you have ideas but ideas have you right. Which and and then we know about the phenomenon of group think and there are so many things that constrain us but I think you can examine all that and not be completely owned by the ideas and completely sucked into groupthink and part of the responsibility as a human is to escape that kind of phenomena which isn't, you know, it's you know, it's it's one of the creative tension things. Again, you're constructed by it but you can still observe it and you can think about it and you can make choices about to some level how constrained you are by it and you know, it's useful to do that and but at the same time and it could be by doing that. You know the group in society, you're you're part of it becomes collectively even more interesting. So you know, so the outside observer will think, wow, you know, all these lex is running around with all these really independent ideas have created something even more interesting than in the aggregate. So, so, so I don't know, I'm those are lenses to look at the situation, but that give you some inspiration, but I don't think they're constraint, right? You know, </s><s>[INST] as a small little quirk of history, it seems like you're related to Jordan Peterson, thank you mentioned he's going through some rough stuff. Now, is there some common you can make about the roughness of the human journey ups and downs? [/INST] Well, I became an expert in Benzo withdrawal. Mhm. Like which is you took benzodiazepines and at some point they interact with gabby circuits, you know, the reduced anxiety and do 100 other things. Like, there's actually no known list of everything they do because they interact with so many parts of your body and then once you're on um you habituate to them and you're you're you have a dependency. It's not like you're a drug dependency, we're trying to get high, it's a it's a metabolic dependency. And then if you discontinue them, there's a funny thing called kindling, which is if you stop them and then go, you know, you'll have a horrible for all symptoms if you go back on them at the same level, you won't be stable and that unfortunately happened to him </s><s>[INST] because it's so deeply integrated to all the kinds of systems [/INST] in the body literally changes the size and numbers of neurotransmitter sites in your brain. So there's a there's a process called the Ashton Protocol where you taper it down slowly over two years. Two people go through that goes through unbelievable hell. And what Jordan's went through seemed to be worse because on advice of doctors, you know, we'll stop taking these and take this. It was the disaster. And he got some, you know, it was pretty tough. Um It seems to be doing quite a bit better intellectually. You can see those brain clicking back together. I spent a lot of time with, I've never seen anybody suffered so much. </s><s>[INST] Well his brain is also like this powerhouse, right? So I wonder does the brain that's able to think deeply about the world suffer more to these kinds of drawings like [/INST] I don't know, I've I've watched videos of people going through withdrawal. They all seem to suffer unbelievably. And you know, my heart goes out to everybody and there's some funny math about this. Some doctor said as best you can tell, you know, there's the standard recommendation is don't take them for more than months and then taper over a couple of weeks. Many doctors prescribe them endlessly, which is against the protocol, but it's common, right? And then Something like 75% of people when they taper, its you know, half the people have difficulty but 75% get off, Okay, 20% have severe difficulty and 5% have life threatening difficulty and if you're one of those is really bad and the stories that people have on this is heartbreaking and and tough. </s><s>[INST] So you put some of the fault of the doctors that just not know what the hell they're doing. [/INST] It's hard to say, it's it's one of those commonly prescribed things, like one doctor said, what happens is if you're prescribed them for a reason and then you have a hard time getting off the protocol basically says you're either crazy or dependent and you get kind of pushed into a different treatment regime, your drug drug addict or a psychiatric patient. And so like one doctor said, you know, I prescribed me from 10 years thinking I was helping my patients and I realized I was really harming them and you know, the awareness of that is slowly coming up. Um the fact that they're casually prescribed to people is horrible and it's bloody scary and some people are stepping on them, but they're on them for life. Like once, you know, it's another one of those drugs that but Benzos long range have real impacts on your personality. People talk about the Benzo bubble where you get disassociated from reality and your friends a little bit, it's really </s><s>[INST] terrible. The mind is terrifying. What we're talking about how, how the infinite possibility of fun, but like it's the infinite possibility is suffering too, which is one of the dangers of like expansion of the human mind. It's like I wonder if all the possible experiences that intelligent computer can have, is it mostly fun or is it mostly suffering? So like if you if you brute force expand the set of possibilities, like are you going to run into some trouble in terms of like torture and suffering and so on? Maybe our human brain is just protecting us from much more possible pain and suffering. Maybe the space of pain, it's like much larger than we could possibly imagine [/INST] and that the world is in the balance. You know, all the all the literature on religion and stuff is, you know, the struggle between good and evil is this balance for very finely tuned for reasons that are complicated. That's all, that's the law and philosophical conversations. </s><s>[INST] Uh, speaking of balance, that's complicated. I wonder because we're living through one of the more important moments in human history with this particular virus, it seems like pandemics have at least the ability to kill off most of the human population at their worst. And they're, they're just fascinating because there's so many viruses in this world. There's so many, I mean viruses basically around the world in the sense that uh they'd been around very long time. They're everywhere. They seem to be extremely powerful and they're just in the distributed kind of way. But at the same time they're not intelligent and they're not even living, Do you have a high level thoughts about this virus that uh like in terms of you being fascinated terrified or someone between [/INST] So I believe in frameworks right? So like one of them is the evolution like we're evolved creatures, right? Yes. And one of the things about evolution is it's hyper competitive and it's not competitive out of a sense of evil. It's competitive sense of there's endless variations and variations that work better when and then over time there's so many levels of that competition. You know like multi cellular life partly exists because of you know the competition between you know different kinds of life forms and we know sex partly exists to scramble our genes so that we have you know genetic variation against uh the invasion of the bacteria and the viruses. And it's endless Like I've had some funny statistic like the density of viruses and bacteria in the ocean is really high and one third of the bacteria die every day because the virus has invaded them. Like 1/3 of </s><s>[INST] them. Like [/INST] like I don't know if that number is true but it was like like there's like the amount of competition and what's going on is stunning and there's a theory as we age we slowly accumulate bacterias and viruses and as our immune system kind of goes down you know that's what slowly kills us and </s><s>[INST] it just feels so peaceful from a human perspective when we sit back and they're able to have a relaxed conversation. Uh, and and there's wars going [/INST] on out there right now you're harboring how many bacteria and you know, the ones, many of them are parasites on you and some of them are helpful and some of our modifying your behavior. And some of them are. No, it's just really, it's really wild. But you know, this particular manifestation is unusual, you know, in the demographic, how it hit in the political, you know, response that engendered and, you know, the healthcare responsive engendered and the technology is gendered. It's kind of wild. </s><s>[INST] The communication on twitter that it uh, every level, all that kind of stuff that ever every single level. Yeah. [/INST] But what usually kills life, The big extinctions are caused by Meteors and volcanoes. </s><s>[INST] That's the one you're worried about supposed to human created bombs that we [/INST] Solar Flares are another good one. You know, occasionally solar flares at the planet. </s><s>[INST] So its nature? [/INST] Yeah, it's all pretty well </s><s>[INST] on another historic moment. This is perhaps outside, but perhaps within your space of framework. So you think about that just happened I guess a couple of weeks ago is um, I don't know if you're paying attention at all as the, the game stop and we'll Wall Street bets. Uh, so it's really fascinating. There's kind of a theme to this conversation today because it's like neural networks. It's cool how there's a large number of people in a distributed way, almost having a kind of fun, we're able to take on the powerful elites, uh elite hedge funds, centralized powers and overpower them. Do you have? Well, I mean, [/INST] I don't know enough about finance, but it was like the Ellen, you know, Robin the guy when they talked, </s><s>[INST] what did you think about that? [/INST] Well, Robinson guy didn't know how to finance system marked, that was clear, right? He was treating like the people who settled the transactions as a black box and suddenly somebody called him up and say, hey black box calling you Your transaction volume means you need to put up $3 billion $3 billion. Like I don't even make any money on these trades, why do I owe $3 billion dollars or you're sponsoring a trade? So, so there was a set of abstractions that, you know, I don't think either. Like, like now you understand it like this happens in chip design, like you buy wafers from TSMC or Samsung or intel and you know, they say it works like this and you do your design based on that and then chip comes back and doesn't work and then suddenly you started having to open the black box is like the transistors really work like they said, you know what's the real issue. So, so it was a, there's a whole set of things that created this opportunity and somebody spotted it now, people spot these kinds of opportunities all the time. So it's been flash crashes, there has been, you know, there's always short squeezes are fairly regular, every ceo I know hates the shorts because they're they're manipulating, they're trying to manipulate their stock in a way that they make money and you know, deprived value from both the, you know, the company and the investors. So the fact that, you know, some of these stocks were so short, it's hilarious that this hasn't happened before. I don't know why, I don't actually know why some serious hedge funds didn't do it to other hedge funds. And some of the hedge funds actually made a lot of money on this. So My guess is we know 5% of what really happened. and a lot of the players don't know what happened. And people probably made the most money, aren't the people that they're talking about. </s><s>[INST] Yeah. Do you think there was something uh I mean, this is this is the cool kind of uh Ellen, uh you're the same kind of conversationalist, which is like first principles questions of like, what the hell happened? Just very basic questions of like, was there something shady going on? Uh what, you know, who are the parties involved is the basic questions everybody wants to know about. [/INST] Yeah. So like, we're in a very hyper competitive world, right? But transactions like buying selling stock as a trust event, you know, I trust the company represented themselves properly, You know, I bought the stock because I think it's going to go up. I trust that the regulations are solid now inside of that. There's all kinds of places where, you know, humans over trust and you know, this, this exposed, let's say, some weak points in the system. I don't know if it's gonna get corrected. I don't know if uh, I don't know if we have close to the real story. You know, my, my suspicion as we don't and listen to that guy, he was like a little wide eyed about and then he did this and then they did that and it was like, I think you should know more about your business than that. But again, there's many businesses when like this layer is really stable. You stop paying attention to it. You pay attention to the stuff that's bugging you or new. You don't pay attention to the stuff that just seems to work all the time. You just, you know, sky's blue everyday California and what, what's, well the range and I was like, what do we do? Somebody go bring in the lawn furniture, You know, like it's getting wet, We don't know why it's getting </s><s>[INST] wet because I was blue for 100 days and now it's, [/INST] you know, so, </s><s>[INST] but part of the problem here with vlad, the, the ceo of Robin hood is the scaling, is that what we've been talking about is there's a lot of unexpected things that happened with the scaling and you have to be, I think the scaling forces you to the and return to the fundamentals. [/INST] Well, it's interesting because when you buy and sell stocks, the scaling is, you know, the stock can only move in a certain range and if you buy a stock, you can only lose that amount of money on the short short market, you can lose a lot more than you can benefit. Like it has a, it has a weird cause, you know, cost function or whatever the right word for that is. So he was trading in a market where he wasn't actually capitalized for the downside. If it got outside a certain range now whether something that various has happened. I have no idea. But at some point the financial risks, both him and his customers was way outside of his financial capacity and his understanding how the system worked was clearly week or or he didn't represent himself. I, you know, I don't know the person and when I listened to him nick, it could have been the surprise question was like, and then these guys called him, you know, it sounded like he was treating stuff as a black box. Maybe he shouldn't have, but maybe his whole pilots for somewhere else and it was going on. I don't, I don't know, </s><s>[INST] I mean, this is, uh, this is one of the qualities of good leaders under fire. You have to perform. And that means to think clearly and to speak clearly. And he dropped the ball on those things because, and I understand the problem, quickly learn and understand the problem. Like at this like basic level, like what the hell happened? [/INST] And my guess is, you know, at some level it was amateurs trading against, you know, experts slash insiders slash people with special </s><s>[INST] information, outsiders and insiders. [/INST] Yeah. And the Insiders, you know, my guess is the next time this happens, we'll make money on it. </s><s>[INST] The Insiders always went [/INST] Well, they have more tools and more incentive. I mean, this always happens like the outsiders are doing this for fun. The insiders are doing this 24/7 </s><s>[INST] but their numbers in the outsiders. This is the interesting thing is [/INST] there's numbers on the Insiders too. Like, like the kind of numbers kind of numbers, </s><s>[INST] but this could be a new era because I don't know. At least I didn't expect that a bunch of editors could, you know, there's, uh, you know, millions of people get [/INST] To go to the next one will be a surprise. </s><s>[INST] But don't you think the crowd that people are planning the next attack? We'll [/INST] see. It has to be surprised. Can't be the same game. So it could be, there's a very large number of games to play and they can be actual about it. I don't know. I'm not an </s><s>[INST] expert, right? That's a good question how the space of games, how how restricted is [/INST] it? And the system is so complicated. It could be relatively unrestricted. And also like, you know, during the last couple of financial crashes, you know, what set it off was, you know, sets of derivative events where, you know, you know, that seemed talibs, you know, the thing is, you're trying to lower volatility in the short run by creating tail events and systems always evolved towards that. And then he always crashed like the gas curve is the, you know, star low ramp plateau crash. It's 100 </s><s>[INST] effective in the long run. Let me ask you some advice to put on your profound hat. What? Uh, there's a bunch of young folks to listen to this thing for no good reason whatsoever. Undergraduate students may be high school students, maybe just young folks are young at heart looking for, uh, the next steps to take in life. What advice would you give to a young person today about life, maybe career, but also life in general, [/INST] get good at some stuff. We'll get to know yourself right. Get good at something that you're actually interested in. You have to love what you're doing to get good at it. You really got to find that. Don't waste all your time doing stuff that's just boring or bland or numbing, right? Don't let old people screw </s><s>[INST] you. [/INST] Well, people get talked into doing all kinds of ship and racking up huge student, you know, student debts and like there's so much crap going on. You know, </s><s>[INST] and the drains your time and drains, [/INST] you know, thesis that, you know, the older generation will like, oh, they're trapping all the young people. </s><s>[INST] And I think there's some truth [/INST] to that just because you're old doesn't mean you stop thinking, I know lots of really original old people. I'm an old person. So, um, but you have to be conscious about it. You couldn't fall into ruts and then do that. I mean when I hear young people spouting opinions, that sounds like they come from Fox news or CNN I think they've been captured by group thinking memes </s><s>[INST] and I also think on their own, [/INST] You know, So if you find yourself repeating what everybody else is saying, you're not gonna have a good life, like, like that's not how the world works and maybe it seems safe, but it puts you at great jeopardy for well being boring or unhappy </s><s>[INST] or how long did it take you to find the thing that you have fun with. [/INST] I don't know, I've been a fun person since I was pretty little So everything I've gone through a couple periods of depression in my life </s><s>[INST] for good reason or for the reason that doesn't make any sense. [/INST] Mm hmm. Like some, some things are hard, Like you go through mental transitions in high school I was really depressed for a year and I think I had my first midlife crisis at 26, I kind of thought is this all there is like I was working at a job that I loved and but I was going to work and all my time is consumed. </s><s>[INST] What's what's the escape out of that depression? What's the answer to is is this all there is [/INST] a friend of mine, I asked him because he was working his ASAP, I said what's your work life balance like like there's you know work friends, family, personal time, You balanced in the net and he said work 80% family 20% and I tried to, I try to find some time to sleep like there's no personal time, there's no passion at that time. Like you know young people are often passionate about work so I was certainly like that but you need to, you need to have some space in your life for different things </s><s>[INST] and that's, that creates uh that makes you resistant to the whole, the dip in the deep dips into depression kind of thing. [/INST] Yeah well you have to get to know yourself to meditation helps some physical, something physically intense helps </s><s>[INST] like the weird places your mind goes kind of thing [/INST] like and why does it happen? What do you do, what you do </s><s>[INST] triggers like the things that cause your mind to go to different places kind of thing or uh your, your [/INST] upbringing for better or worse whether your parents are great people or not, you you come into adulthood with all kinds of emotional burdens and you can see some people are so bloody stiff and restrained and they think, you know, the world's fundamentally negative. Like you maybe you have unexplored territory. </s><s>[INST] Yeah. [/INST] Or you're afraid of something </s><s>[INST] definitely afraid of quite a few things. [/INST] You got to go face him. Like what's the worst thing that happened? You're gonna die, right? Like that's inevitable. You might as well get over that like, 100%. That's right. People are worried about the virus, but you know, the human condition is pretty deadly. </s><s>[INST] There's something about embarrassment that I've competed a lot of my life and I think that if I'm to introspect it, the thing I'm most afraid of is being like humiliated. I [/INST] think nobody cares about your the only person on the planet that cares about you being humiliated so I can't really useless thought it is. It's like </s><s>[INST] uh [/INST] you're humiliating something happened in a room full of people and they walk out and they didn't think about it one more second. Or maybe somebody told a funny story to somebody else and then </s><s>[INST] interrogated throughout. Yeah. [/INST] Yeah. No, I know it too. I mean, I've been really embarrassed about it that nobody cared about myself. Yeah, it's a funny </s><s>[INST] thing. So the worst thing ultimately is just uh [/INST] yeah, but that's a cage and then you have to get out of it, like, once you, here's the thing, once you find something like that, you have to be determined to break it because otherwise you'll just, you know, so you accumulate that kind of junk and then you die as a, you know, a mess. </s><s>[INST] So the goal, I guess it's, it's like occasionally in a cage, I guess the goal is to die in the biggest possible cage. [/INST] Well, ideally you'd have no cage. Mhm. You know, people do get enlightened, I've got a few, it's great. </s><s>[INST] You found a few, there's a few out there, I don't know, of course. Uh [/INST] huh. Either that or they, you know, it's a great sales pitch that was like enlightened people write books and doing all kinds of </s><s>[INST] stuff. It's a good way to sell a book, I'll give you that. [/INST] You've never met somebody, you just thought they just kill me, like they just like mental clarity, humor. </s><s>[INST] No, 100%, but I just feel like they're living in a bigger cage. They [/INST] have their own, they don't think there's a cage and they're still kids, you secretly suspect there's always a case. </s><s>[INST] Uh There's no there's nothing outside the [/INST] university outside the cage. Uh </s><s>[INST] you work, you work, you worked in a bunch of companies, you let a lot of amazing teams um I don't I'm not sure if you've ever been like in the early stages of a startup, but do you have advice for uh somebody that wants to do a startup, or build a company, like, build a strong team of engineers that are passionate just want to uh solve a big problem. Like is there uh more specifically on that point, [/INST] you have to be really good at stuff, you're going to lead and build a team, you better be really interested in how people work and think </s><s>[INST] the people or the solution to the problem. So there's two things, right? One is how people work, and the other is [/INST] actually there's quite a few successful startups. So it's pretty clear the founders don't know anything about people, like, the idea was so powerful that problem, but I suspect somewhere early, they hired some people who understood people because people really need a lot of care and feeding to collaborate and work together and feel engaged and work hard, you know, like startups, that was all about out producing other people like your nimble, because you don't have any legacy, you don't have, you know, a bunch of people who are depressed about life, you know, just showing up, you know, So startups have a lot of advantages that way. Do </s><s>[INST] you like the steve jobs, talked about this idea of eight players and the players? I don't know if you uh know this formulation? [/INST] No. Um did that organizations that get taken over by PB player leaders often really underperformed the heresy players that said in big organizations, there's so much work to do, like, and there's so many people are happy to do what you know like the leadership or the big idea. People who can consider menial jobs and you know you need a place for them but you need an organization that both values and rewards them but doesn't let them take over the leadership of it. </s><s>[INST] Got it. But so so you need to have an organization that's resistant to that. But in the early days the notion with steve was that like one B player in the room of a players will be like destructive to the whole. [/INST] I've seen that happen. I don't know if it's like always true like you know you run into people who clearly be players but they think they're very players and so they have a loud voice at the table and they make lots of demands for that. But there's other people are like I know I am, I just want to work with you know, cool people on cool ship and just tell me what to do and I'll go get it done. You know? So you have to again, this is like people skills like what kind of person is it? You know, I've been some really great people I love working with that weren't the biggest I'd people they're most productive ever, but they show up, they get it done. You know, they create connection and community that people value. It's it's pretty diverse. I don't think there's a recipe for that. </s><s>[INST] I got to ask you about love, [/INST] I heard you into this now </s><s>[INST] into this love [/INST] thing is this is you think this is your solution to your depression? </s><s>[INST] No, I'm just trying to like you said the enlightened people on occasion trying to sell a book. I'm writing a book about [/INST] love your writing a book about </s><s>[INST] No, I'm not, I'm [/INST] not a friend of mine got a wee said you should really write a book about your, your management philosophy. Said it would be a short book. Uh </s><s>[INST] well that one was pretty well uh what role do you think? Love family, friendship, all that kind of uh human stuff. Playing a successful life. You've been exceptionally successful in the space of like running teams, building cool ship in this world, creating some amazing things. What did love getting away? Did love help the family? Getting away to family help [/INST] friendship? You want the engineers answer please. So first love is functional, right? </s><s>[INST] It's functional in that way. [/INST] So we habituate ourselves to environment and actually Jordan told me Jordan Peterson told me this line, so you go through life and you just get used to everything except for the things you love. They remain new. Like this is really useful for, you know, like, like other people's Children and dogs and trees, you just don't pay that much attention to your own kids, you're monitoring really closely like and if they go off a little bit because you love them. If you're smart, if you're going to be a successful parent, you notice it right away, you don't habituate just things you love and if you want to be successful at work, if you don't love it, you're not going to put the time and somebody else that's something else that loves it, like, because it's new and interesting and that lets you go to the next level. </s><s>[INST] So the thing is just a function that generates newness and novelty and surprise is, you know, all those guys, [/INST] it's really interesting, like there's people figured out lots of frameworks for this, you know, like, like humans seem to go in partnership, goes through, you know, interest, like somebody, suddenly somebody is interesting and then you're infatuated with him and then you're in love with him and then you, you know, different people have ideas about parental love or mature love, like you go through a cycle of that, which keeps us together and it's, you know, super functional for creating families and creating communities and making you support somebody despite the fact that you don't love them, like, and and it can be really enriching, you know, No. Now in the work life balance scheme, if all you do is work, you think you may be optimizing your work potential, but if you don't love your work or you don't have family and friends and things, you care about your brain isn't well balanced, like everybody knows the experience of you worked on something all week. He went home and took two days off and you came back in the odds of you working on the thing you're picking up right where you left off a zero. Your brain re factored it. Mhm. For being in love is great. It's like changes the color of the light in the room. It creates a spaciousness. That's that's different. It helps you think </s><s>[INST] it makes you strong? Bukowski had this line about love being a fog that dissipates with the first light of reality in the morning. That's [/INST] depressing. I think it's the other way around </s><s>[INST] it lasts what you like you said is there's a function. It's the thing that [/INST] you can be the light that actually enlivens your world and creates the interest and the power and the strength and the to go do something. It's like, like, like that sounds like, you know, there's like physical of emotional, of intellectual love spiritual, </s><s>[INST] right? Isn't all the same thing kind of, [/INST] you should differentiate that maybe that's your problem in your book. You should you should </s><s>[INST] refine that a little bit different [/INST] chapters. Yeah. There's different chapters. </s><s>[INST] What's that? What's these are, aren't these are just different layers of the same thing with the stack physical [/INST] people people, some people are addicted to physical of and they have no idea about emotional or intellectual love. Mhm </s><s>[INST] I don't know if they're the same thing. [/INST] So, I think they're different. </s><s>[INST] That's true. That could be different. It would be, I guess the ultimate goal is for it to be the [/INST] same. If you want something to be bigger and interesting, you should find all its components and differentiate them climate together. People do this all the time. They yeah, the modularity get your abstraction layers, right? And then you can you have room to breathe </s><s>[INST] or maybe you can write the forward to my book about [/INST] love or the afterwards and after we really tried. I feel like this has been a lot of progress of this book. Uh well, you have things in your life that you love. </s><s>[INST] Yeah, [/INST] yeah. So, </s><s>[INST] and then they are, you're right, they're modular. It's [/INST] it's well, you can have multiple things with the same person or the same thing. And but </s><s>[INST] yeah, depending on the moment of the day, [/INST] like what Penkovsky described as that moment you go from being in love to having a different kind of love, right? And that's the transition. But when it happens, if you'd read the owner's manual and you believe that you have said, oh, this happened, it doesn't mean it's not love, it's a different kind of love, but mhm But maybe there's something better about that as you grow old, if all you do is regret how you used to be. It's sad, right? You should have learned a lot of things because like who you can be in your future self this it's actually more interesting than possibly delightful than you know being a mad kid in love with the the next person. Like that's super fun when it happens. But That's that's you know 5% of the possibility. </s><s>[INST] But yeah that's right that there's a lot more fun to be had in a long lasting stuff [/INST] or meaning you know if </s><s>[INST] that's meeting which is a kind of fun, the deeper kind of fun [/INST] and it's surprising you know that's like the thing I like surprises you know and you just never know what's gonna happen but you have to look carefully, you have to work at it, you have to think about it you know it's </s><s>[INST] Yeah you have to see the surprises when that happened right? You have to be looking for from the branching perspective you mention regrets. Uh do you have regrets about your own trajectory? [/INST] Oh yeah, of course. Yeah. Some of it's painful but you wanna hear the painful stuff I would say like in terms of working with people when people did say stuff I didn't like especially if it was a bit nefarious. I took it personally and I also I felt it was personal about them but a lot of times like humans are, most humans are a mess right and then they act out and they do stuff and this psychologist I heard a long time ago said you tend to think somebody does something to you but really what they're doing is they're doing what they're doing while they're in front of you. It's not that much about you. Yeah. Right. And as I got more interested in, you know when I work with people I think about them and probably analyze them and understand them a little bit and then when they do stuff I'm way less surprised and I'm waiting. You know, and if it's bad I'm way less hurt and I react way less. Like I sort of expect everybody's got their ship. </s><s>[INST] Yeah. It's not about you. [/INST] It's about me that much. It's like, you know, you know, you do something and you think you're embarrassed but nobody cares like and somebody's really mad at you at the odds of it being about, you know, they're getting mad the way they're doing that because of some pattern they learned and you know, and maybe you can help them if you care enough about it. But or you can step you can see it coming and step out of the way like like I wish I was way better at that. I'm a bit of a hothead </s><s>[INST] and support. You said it was steve that was a feature not a bug. [/INST] Yeah. Well he was using it as the counter for so orderliness that would crush </s><s>[INST] his work. You were doing the same. [/INST] Maybe I don't think I don't think my my vision was big enough. It was more like I just got piste off and did stuff. </s><s>[INST] I'm sure that's what [/INST] you're telling me. I don't know if it had, it didn't have the amazing effect of creating a trillion dollar company. It was more like I just got piste off and left and or made enemies that he shouldn't have been. Yeah, it's hard. Like I didn't really understand politics until I worked at Apple where you know, steve was a master player of politics and his staff had to be or they wouldn't survive him. And it was definitely part of the culture. And then I've been in companies where they say it's political, but it's all, you know, fun and games compared to Apple and it's not that the people at Apple are bad people is just they operated politically at a higher level. You know, it's not like, oh somebody said something bad about somebody, somebody else which is most politics. It's, you know, they had strategies about accomplishing their goals sometimes, you know over the dead bodies of their enemies, you know, with sophistication and yeah, more game of Thrones. Sophistication and like a big time factor rather than a, you know, </s><s>[INST] well that requires a lot of control over your emotions. I think uh to have a bigger strategy in the way behave and it's it's [/INST] effective in the sense that coordinating thousands of people to do really hard things where many of the people in there, don't understand themselves, much less how they're participating, creates all kinds of, you know, drama and problems that, you know, our solution is political in nature. Like how do you convince people, how do you leverage them? How do you motivate them? How do you get rid of them? How do you know, like there's there's so many layers of that that are interesting. And even though some some of it, let's say maybe tough, it's not evil unless, you know, you use that skill to evil purposes, which some people obviously do. It's a skill set that operates and I wish I'd, you know, I was interested in it, but you know, it was sort of like I'm an engineer, I do my thing and you know, there's times when I could have had way bigger impact if I knew how to if I paid more attention and knew more about that, </s><s>[INST] about the human layer, this [/INST] that yeah, that human political power expression layer of the stock just complicated and there's lots to know about it. I mean, people are good at it are just amazing when they're good at it. And let's say relatively kind and oriented a good direction. You can really feel you can get lots of stuff done and coordinate things you never thought possible. All right. But all people like that also have some pretty hard edges because you know, it's a heavy left and I wish I'd spent more time at that when I was younger, But maybe I wasn't ready, you know, I was wide eyed kid for 30 years, </s><s>[INST] it's a little bit of a kid, you know, what do you hope your legacy is when there's a, there's a book, like a Hitchhiker's Guide to the Galaxy and this is like a one sentence of entry ball jim color from like that guy lived at some point. There's not many, not many people would be remembered. You're one of the sparkling little human creatures that had a big impact in the world. How do you hope you'll [/INST] be remembered? My daughter was trying to get, She added my Wikipedia page to say that I was a legend in the guru, but they took it out. So she put it back in. She's 15. Think, I think that was probably the best part of my legacy. She got her sister and they were all excited. They were trying to put it in the references because there's articles since that on the top of </s><s>[INST] that. So the eyes of your kids, your legend. [/INST] Well they're pretty skeptical because they don't be better than that. They're like dad. So yeah, that's, that's that kind of stuff is super fun in terms of the big legend stuff quicker. Okay, Legacy. I don't really </s><s>[INST] care. There's just an engineer. [/INST] Yeah, I've been thinking about building a big pyramid. So either debate with a friend about whether pyramids or craters or cooler and you realize that There's graders everywhere, but you know, they built a couple of pyramids 5000 years ago </s><s>[INST] and I remember [/INST] you were still talking about </s><s>[INST] it. I [/INST] think that would be cool. </s><s>[INST] Uh, those aren't easy to [/INST] build. Oh, I know. And they don't actually know how they built them, which is great. </s><s>[INST] Uh, it's either uh, a G. I. Or aliens could be involved. So I think, I think you're gonna have to figure out quite a few more things than just the basics of civil engineering. So I guess you hope your legacy is pyramids. [/INST] That would, that would be cool. And my Wikipedia page, you know, getting updated by my daughter periodically. Like those two things would pretty much make it </s><s>[INST] Jim. It's a huge honor to talking to you again. I hope we talked many more times in the future. I can't wait to see what you do with 10 story. I can't wait to use it. I can't wait for you to revolutionize yet another space in computing. It's a huge honor to talk to you. Thanks for talking today. Thanks for listening to this conversation with Jim Keller and thank you to our sponsors athletic greens, all in one nutrition drink Brooklyn and sheets express VPN and belcampo grass fed meat, click the sponsor links to get a discount and to support this podcast and now let me leave you with some words from Alan turing those who can imagine anything can create the impossible. Thank you for listening and hope to see you next time. [/INST] </s> |
<s>[INST] Isn't that a few fundamental? If you mind, why do you say that tech billionaires don't have a few money [/INST] when you need to hire private security to protect your family? How do you protect your two Children? I don't have those yet bingo. Yeah. My point is is that f you money insulate everything that you care about, it's not just about you. </s><s>[INST] So you're saying as the level of responsibility grows. The amount of money required for a few. [/INST] We have a war going on. The war is on academic freedom. Academic freedom used to be present in the system as a in terms of the idea we we trust our elite. Now we have an idea like you want to be the elite. You know, you want to lord above us. That's like, first of all, there's like a populist anti elitist thing. Then there's the idea that, um, we're gonna defer tenure for forever. Then we're gonna tell people stay in your lane. Your tenure is only good for your own particular tiny micro subject. And then we're going to also control your grants and we'll be able to load up your teaching load if we don't like who you are and will make your life absolutely impossible. We lost academic freedom and we ushered in peer review, which was a disaster. And then we lost funding so that people were confident that they would have the ability to do research no matter what they said. And as a result, what you find is is a world in which there is no ability to get people to say, no, I'm not going to sign your diversity and inclusion forced loyalty oath. I won't sign any loyalty oath, Get the hell out of my office. </s><s>[INST] Uh, you a few. And you're connecting money to that. But [/INST] my point is is that academic freedom is the, the whole idea behind it was that you will have the freedom of a billionaire on a much smaller seller. Okay, we've lost that. Yeah. The only reason in part that I wanted to go into academic academics as a profession as opposed to wanting to do physical or mathematical research. The great prize was freedom ralph Gomory of the Sloan Foundation previously of IBM Research pointed at, he says if you lose freedom, you lose the only thing we had to offer top minds, top minds value their intellectual freedom and their physical and economic security at a different level than other human beings. And so people say, you know, I understand dude, and you have the ability to do X, y and Z. What's the problem? It's like, well, I value my ability to raise the middle finger as an american practically above everything else. </s><s>[INST] I want to talk to you about freedom here in the context of something you've mentioned, which one way to take away freedom is to put human being into a cage to create constraints. The other one that worries me is something I think you've spoken to to twitter a little bit on twitter is we bleed freedom by kind of slowly mhm uh scaring you into not doing not expressing the full spectrum of opportunities. You can his freedom. So like when you ban donald trump, when you ban parler, you give a little doubt in the minds of millions like me, a person who is a tech person who's an entrepreneur, there's a little, that's what I'm afraid of when I look in the mirror, is there now a little doubt in there, that limits the amount of options I will try. [/INST] How certain are you that the Covid virus didn't come from the Wuhan lab And his bio safety level four, we both know that we're both supposed to robotically say the idea that the Covid virus came from a lab is a discredited conspiracy theory. There is no evidence that suggests that this is true. The World Health Organization and the CDC have both upon this. To say otherwise would be incredibly irresponsible. </s><s>[INST] And the threat of that is the thing that ultimately limits the freedoms. We feel [/INST] I should be tweeting about Jeffrey Epstein all the time? </s><s>[INST] And you're afraid? It's also, [/INST] I mean, I said it in the public, why is it? We don't ask where the records are, from Villard House, where the financial records, where the sec filings, where are the questions on, on the record to the intelligence uh agencies? Was he known to be part of the intelligence community? So we're not interested in asking questions like, am I going to die as a result of asking the question was Jeff Epstein part of the intelligence community of any nation, Is there a reason we're not asking about the financial records of the supposed hedge fund that he didn't run? Just like the Wuhan lab. </s><s>[INST] Okay, how do we get to the core of the Jeffrey Epstein, the truth behind Jeffrey Epstein in a sense. I mean, there's, there's some things that are just like useless conspiracy theories around it, even if they're true, there's some things that get you hate [/INST] to say it, you're not gonna like it. Look at the 1971 media pennsylvania break in of the citizens committee to investigate the FBI those kids. And by the way, they weren't all kids did what had to be done. They broke in, They broke the law. It was an incredible act of civil disobedience and God bless judy fine goal for taking to her. She was going to take to her grave the cheap part of this. Like the coolest thing of all time. They didn't say anything for forever. </s><s>[INST] So civil disobedience, I mean, you have to, we [/INST] are founded on civil disobedience. Civil disobedience is incredibly, you screw it up and you're just a vandal, You screw it up. You're a hooligan. Those, those cats were so disciplined. It's an art form and it was an art form and they risked everything. They were willing to pay with their freedom. Those are the sorts of people who earned the right by putting themselves at risk. I would not do this. I am not volunteering to break into anything. I think it was William David in who was a student of Murray gell mann and a physics professor at Haverford who corralled these people and lead this effort and right now, what we need is somebody to blow the lid off of? What is controlling everything we have? I'm happy to hear that. It's a system of incentive structures that it's a system of selective pressures. I'm happy to find out that it's emergent. I'm happy to find that it's partially directed by our own intelligence community. I'm happy to hear that uh in fact we've been penetrated by north Korea, Iran china and Russia, but I need to know why people aren't like the firebombing of the courthouse in Portland Oregon has no explanation. And somehow this is normal. This is not normal to any human being. We have video that people don't believe. And you know, I come back to the Shaggy defense. It wasn't me, you know, it's like um remember that song, </s> |
<s>[INST] Yeah, by the way, I should push back on the billionaires because there's sometimes a negative connotation to the word billionaire. It's not a billionaire, it's a company versus government because governments are billionaires and trillion airs. It's not it's not the wealth, it's the capitalist uh imperative. So uh which I think deserves a lot more praise than people are giving it. I'm I'm troubled by this sort of criticism, like it's billionaires playing with toys for their own pleasure. I think what some of these companies like SpaceX and Blue Origin are doing is some of the most inspiring engineering and even scientific work ever done in human history. [/INST] No, I agree. I think that people have made the greatest wealth are people who have really been mega benefactors. I mean, I think, you know, some </s><s>[INST] of them, some of [/INST] them, yes, some of them, but but but those who who founded google and all that and and even amazon their their their their beneficiaries then they're quite different category in my view from those who just shuffle around money or or or crypto coins and things like that, who are </s><s>[INST] not yet. [/INST] But I think if they use their money in these ways that's fine. But I think it's true that the far more money is owned by us, collectively as taxpayers. But I think the fact is that in a democracy, um there'd be big resistance to exposing human beings to very high risks if in a sense, we share responsibility for it. </s><s>[INST] And [/INST] that's that's the reason I think we've done much more cheaply by by by these private funders. </s><s>[INST] That's an interesting hypothesis, but I have to push back. I don't know if it's obvious why Nasa spends so much money and takes such a long time to develop the things I was doing so before Elon musk came along. Because I would love, I would love to live in a world where government actually uses taxpayer money to get some of the best engineers and scientists in the world and actually work across governments, Russia china United States, european union together to do some of these big projects. It's strange that Ellen is able to do this much cheaper much faster. It could have to be do with risk aversion. You're right, but I think it's [/INST] it's the it's that it's that he had all the the whole assembly within this one building as it were rather than depending on a supply chain. But I think it's also that he had a Silicon valley culture and younger people. Whereas the big aerospace companies Boeing and Lockheed martin, they had people who were left over from the Apollo program in some cases and so they weren't quite so lively and indeed quite apart from the controversial issues of the future of human spaceflight. In terms of the next generation of big rockets. Then the one that musk is going to launch for the first time this year, huge one is going to be far far cheaper than the one that Nasa has been working on the same time. And that's because it will have a reusable first stage and it's going to be great. It can launch over 100 tons into Earth orbit and instead that's going to be make it feasible to do things that I used to think we're crazy like having solar energy from space. That's no longer so crazy if you can do that. And also for science because its nose cone could contained within it, something as big as the entire unfurled James Webb telescope mirror and have a big telescope much more cheaply if you can launch it all in one piece. And so it's going to be hugely beneficial to science and to any practical use of space to have these cheaper rockets that are far more completely reusable anyways Nasa had. So I think mosques are a tremendous service to the space exploration and the whole space technology through these rockets </s><s>[INST] certainly. Plus it's some big sexy rocket. It's just great engineering. [/INST] Of course, yeah. It's </s><s>[INST] like looking at a beautiful big bridge that humans are capable of descendants of apes are capable to do something so majestic. [/INST] And also the way they land coming down at this bar. That's amazing. </s><s>[INST] It's both controls engineering. It's um increasing sort of intelligence in these rockets but also great propulsion engineer materials uh, on your ship. And it just inspires. It just inspires so many people [/INST] know I'm entirely with you on that. </s><s>[INST] So would it be exciting to you to see a human being stepped foot on mars in your lifetime? Yes. [/INST] I think it's unlikely in my lifetime since I'm I'm so ancient but but I think this essentially is going to happen And I think that that will indeed be exciting. And I hope there will be a small community by the end of the century. But as I say, I think they may go with one way tickets or accepting the risk of of of no return. So they've got to be people like that. And I still think it's going to be hard to persuade the public to send people when you say straight out that they may never come back. But of course the Apollo astronauts, they took a high risk And in fact in in my previous book, I quote the speech that has been written for Nixon to be read out if Neil armstrong got stuck on the moon. And it was written by one of his, his advisor's very eloquent speech, you know about how they have come to a noble end et cetera. But of course there was a genuine risk at that time, but that may have been accepted. But clearly the crashes of the shuttle were not acceptable to the american public Even when they were told that this was only a 2% risk given how often they launched it. And so that's what leads me to think that it's got to be left to the kind of people who are prepared to take these risks. And and II think think of America event was a guy called Steve Fossett who was aviator, did all kinds of crazy things, you know, and then the guy who fell super sonically um with the parachute from very high altitude, all these people, we all cheer them on, they extend the bounds of humanity. But I don't think the public would be so happy to fund them. </s> |
<s>[INST] just it was just walking down the hall and nobody paying attention and it's just like walking down this long hall and I'm like looking around is anyone like what's happening here? So presumably some kind of automation was doing the map. I mean the whole environment is probably really well mapped but it was just it gave me a picture of a world where robot is doing doing his thing wearing a couple hats, just going down the hall like getting some coffee or whatever. Like I don't know what it's doing, what's the mission. But I don't know for some reason that really stuck with me. You don't often see robots that aren't part of a demo or that aren't uh you know like with a semi autonomous autonomous vehicle, like directly doing a task. This was just chilling. Yeah. I don't know. [/INST] Well yeah. You know I mean we're at M. I. T. Like when I first got to M. I. T. I was like okay where's all the where's all the robots? And they were all like broken or like not damn going. So. Yeah. And and and what really excites me is that we're about to have that we're about to have so many moving wrote about to. Well it's coming, it's coming in our lifetime that we will just have robots moving around already. Seeing the beginnings of it there's delivery robots and some cities on the sidewalks and I just love seeing like the tick talks of people reacting to that because yeah, you see robot walking down the hall with a cowboy hat. Like what the, what is this? This is awesome and scary and kind of awesome and people either love or hate it. That's one of the things that I think companies are underestimating that people will either love a robot or hate a robot and nothing in between. So it's just again an exciting time to be alive. </s><s>[INST] Yeah. I think kids almost universally at least in my experience love them. Love legged robots. [/INST] If they're not like my, my son hates the remote because ours is loud. </s><s>[INST] Oh, that yeah, no, the legs. The legs [/INST] because </s><s>[INST] your son. Um, do they understand rumored to be a robot? [/INST] Oh yeah. My kids, that's one of the first words they learned. They know how to say beep boop and yes, they think the room as a robot. </s><s>[INST] Do they project intelligence out of the thing? [/INST] Well, we don't really use it around them anymore for the reason that my son is scared of it. </s><s>[INST] Yeah. [/INST] I think they would like even a Roomba because it's moving around on its own. I think kids and animals view it as an agent. </s><s>[INST] So what do you think if we just look at the state of the art of robotics? What do you think robots are actually good at today? So if we look at today, [/INST] you mean physical robots? </s><s>[INST] Yeah, physical robots, [/INST] wow. </s><s>[INST] Like what are you impressed by? So I think a lot of people, I mean that's what your book is about is have maybe a not a perfectly calibrated understanding of where we are in terms of robotics was difficult to robotics. What's easy in robotics? [/INST] Yeah, we're way behind where people think we are. So What's impressive to me. So let's see. 011 thing that came out recently was Amazon has this new warehouse robot and it's the first autonomous warehouse robot that can is safe for people to be around. And so like it's kind of most people, most people I think envision that our warehouses are already fully automated and that there's just like robots doing things. It's actually still really difficult to have robots and people in the same space because it's dangerous for the most part robots, you know, because especially robots that have to be strong enough to move something heavy, for example, they can really hurt somebody. And so until now a lot of the warehouse robots had to just move along like pre existing lines which really restrict what you can do. Um And so having I think that's that's one of the big challenges and one of the big like exciting things that's happening is that we're starting to see more Kobach ticks in industrial spaces like that where people and robots can work side by side and not get harmed. </s><s>[INST] Yeah. That's what people don't realize that the physical manipulation tasks with humans, it's not that the robots want to hurt you. I think that's what people are worried about. Like this malevolent robot gets them out of its own and wants to destroy all humans. Now. It's, you know, it's actually very difficult to know where the human is and to to respond to the human and dynamically and collaborate with them on a task, especially if you're something like an industrial robotic arm, which is extremely powerful. See this some of the some of those arms are pretty impressive now that you can just you can you can grab it, you can move it. So the collaboration between human robot and the factory setting is really fascinating. Um Do you think they'll take our jobs? [/INST] I don't think it's that simple. I think that there's a ton of disruption that's happening and will continue to happen. Um You know, I think speaking specifically of the amazon warehouses, that might be an area where it would be good for robots to take some of the jobs that are, you know, where people are put in a position where it's unsafe and they're treated horribly and you know, probably it would be better if a robot did that and amazon is clearly trying to automate that job away. So I think there's gonna be a lot of disruption, I do think that robots and humans have very different skill sets. So while a robot might take over a task, it's not gonna take over most jobs. Um I think just things will change a lot. Like I don't know one of the examples I have in the book is mining um So there you have this job that is very unsafe and that requires a bunch of workers and put them in unsafe conditions and now you have all these different robotic machines that can help make the job safer. And as a result now people can sit in these like air conditions, remote control stations and like control these autonomous mining trucks. And so that's a much better job but also they're employing less people now. So it's it's just a lot of I think from a bird's eye perspective, you're not going to see job loss, you're going to see more jobs created because that's I think the future is not robots just becoming like people and taking their jobs. The future is really a combination of our skills and then the supplemental skill set that robots have to increase productivity to help people have better, safer jobs too. Give people work that they actually enjoy doing and are good at. Um But it's really easy to say that from a bird's eye perspective and um ignore kind of the the rubble on the ground as we go through these transitions because of course specific jobs are going to get lost </s> |
<s>[INST] There's an interesting idea, I think Michael malice borrowed up as a test whether you're on the left or the right. Uh, the question he asks, which is, do you think some people are better than others, if you say yes, he claims you're on the right? If you start answering, if you start like saying a lot of things like uh you're on the left. So if you start explaining yourself well equivocating and it's a good term for it. I was really so in this, in this test, I suppose I would be on the left because I'm uncomfortable with the idea that some people are better than others as a basic feeling as a starting point in the way you think about the world. Because as we're talking about everybody is a hero of their own story. When you start to think some people are better than others as a starting axiom, it's like a slippery slope to where you think you're way better than others. And then you start to like basically it's ok to take advantage of a large percent of the population for the greater good, totally. And then you go into Stalin mode and Hitler mode or it's okay to murder a large part of the population for the greater good. So it's like, it's this very dangerous, slippery slope of my mind. So I try to not uh yeah, I was always uncomfortable that kind of test or even that kind of thought. And you have the same applies. And suppose in uh in government and central banking is if you think some people are better than others, applying your idea what is good can have large scale detrimental effects. [/INST] Of course, I'm glad you didn't pose me the </s><s>[INST] question? I [/INST] mean I think it may be not the left right axiom isn't the disjunction is in the way I would sort of put it. But you know to me it's just if you reason in a consequentialist way, you know, that lends itself to authoritarianism whereby you think you can shape society and only you can shape society in a positive direction according to your, you know specific objectives. </s><s>[INST] So let's step onto the land of sturdiness. That is Bitcoin. What is Bitcoin And in your view, what are you know the principles, the philosophical foundations of Bitcoin? [/INST] Well they coined the term I think refers to two things specifically. So one is the protocol for conveying value through communications channel. So just a set of rules that we collectively opt into in order to transact online or just at a distance. And then the other thing is the name of the asset, the sort of monetary unit which circulates within the system. And that always confused people a lot because it's like well uppercase Bitcoin lowercase Bitcoin. Why didn't Satoshi just give them different names like in ethereum you've got a theory um the system and then either, although people don't really talk about ether very much but they you know chose to distinguish them in Bitcoin for whatever reason they're not distinct? So the two Bitcoins get co mingled all the time in the explanations. </s><s>[INST] Did you find that the problem that confuses things? I mean what's what's really a distinction between the protocol and the currency? [/INST] Well they are sometimes distinguished practically like you can transact with Bitcoin outside of the Bitcoin protocol for instance. Right? So you know you can transact with Bitcoin on ethereum or have Bitcoin on an open dime here. This would be a Bitcoin transaction, it wouldn't settle on the Bitcoin network. </s><s>[INST] Do you mind explaining what you have on the table before us? [/INST] Yeah so I brought you some presents. That's awesome. This isn't a bribe. This is just a proof of concept. So this is basically um a bit coin bearer instrument so I put 100 bucks a bic went on here and to spend it. You have to basically physically destroy part of the device used to poke a hole. You know poke off one of the little transistors on this so it can only be spent once. So and you can extract the private key from this device. So the private key was generated on device always stays on the device so what it means without like breaking off like a small part. So this basically is a way to physically and stan she ate Bitcoin so it's </s><s>[INST] um it's kind of [/INST] old. Yeah effectively. So here </s><s>[INST] this one's limited edition, it's orange. So what is it called [/INST] again? Open dime? The point is if you wanted to settle a Bitcoin transaction instantly. The kind of the same way that cash transaction is instant final settlement right? You would do it with a device like this. So if I was buying a house from your you know you might prefer to do it with a physical bearer instrument as opposed to waiting for confirmation on the Bitcoin Blockchain. So the moment I hand that over to you goes in your possession you're the owner. There's no way for me to have retained the private key. Like I could have created a Bitcoin paper wallet and given that to you but you have no assurance that I didn't copy down that you know the key elsewhere. So this solves that problem </s><s>[INST] because this is a physical sensation of the Bitcoin ah transaction outside the Bitcoin protocol. That's right. This is your transacting the currency outside of the [/INST] protocol. It's analog. Bitcoin. We're running an analog which I was like because Bitcoin is this immaterial thing. And so it's nice to have physical totems. </s><s>[INST] How much does it cost to manufacture this? Dino [/INST] Like $15 or something </s> |
<s>[INST] And that's especially important for that made most clear in reinforcement learning where simulated environment is uh you can only learn as much as you can simulate. And that's what made uh Deepmind made very clear what the other aspect of the environment, which is the self play mechanism of uh the other agent of the competitive behavior, which the other agent becomes the environment essentially. And that's I mean one of the most exciting ideas in ai is the self play mechanism that's able to learn successfully. So there you go. There's a there's a thing where competition is essential for learning at least in that context. So if we can step back into another sort of beautiful world, which is the actual mechanics, the dirty [/INST] mess </s><s>[INST] of it of the human brain, is there something for people who might not know? Uh is there something you can comment on or describe? The key parts of the brain that are important for intelligence or just in general? What are the different parts of the brain that you're curious about that you've studied and uh that are just good to know about when you're thinking about cognition. [/INST] Well, my area of expertise, if I have one is the prefrontal cortex. So uh what's that </s><s>[INST] or do we? It depends on [/INST] who you ask. The technical definition is is anatomical. There are there are uh parts of your brain that are responsible for motor behavior and they're very easy to identify. Um and um the region of your cerebral cortex, they need to sort of outer crust of your brain that lies in front of those is defined as the prefrontal cortex. </s><s>[INST] And we say anatomical, sorry to interrupt. So that's referring to uh sort of the geographic region as opposed to some kind of functional definition. [/INST] Exactly. So that this is kind of the coward's way out. I'm telling what the prefrontal cortex is just in terms of like what part of the real estate it occupies. The thing in the front of the exactly. And and in fact the early history of uh you know the neuro scientific um investigation of what this like front part of the brain does is sort of funny to read because uh you know. Mhm. It was really it was really world War One that started people down this road of trying to figure out what different parts of the brain the human brain do in the sense that uh there were a lot of people with brain damage who came back from the war with brain damage. And it that provided as tragic as that was. It provided an opportunity for scientists to try to identify the functions of different brain regions and that was actually incredibly productive. But one of the frustrations that neuropsychologist faced was they couldn't really identify exactly what the deficit was um that arose from damage to this these most you know kind of frontal parts of the brain. It was just a very difficult thing to you know to um you know, to pin down, there were a couple of neuro psychologists who identified through through a large amount of clinical experience and close observation. They started to um put their finger on a syndrome that was associated with frontal damage. Actually one of them was a Russian uh neuropsychologist named Gloria, who you know, students of cognitive psychology still read. Um and what he started to figure out was that the frontal cortex was somehow involved in flexibility. The in in in guiding behaviors that required someone to override a habit or to do something unusual or to change what they were doing in a very flexible way from one moment to another. </s><s>[INST] So focused on like new experiences. So uh so the way your brain processes and acts in new experiences. Yeah. [/INST] What later helped bring this function into better focus was a distinction between controlled and automatic behavior or to in other literatures, this is referred to as habitual behavior versus goal directed behavior. So it's very very clear that the human brain has pathways that are dedicated to habits, to things that you do all the time and they need to be automatic. Ized so that they don't require you to concentrate too much. So that leaves your cognitive capacity for you to do other things. Um Just think about the difference between uh driving when you're learning to drive versus driving after you're fairly expert. There are brain pathways that slowly absorb those frequently uh performed behaviors so that they can be habits so that they can be </s><s>[INST] automatic. So that that's kind of like the purest form of learning, I guess it's happening there, which is why I mean this is kind of jumping ahead, which is why that perhaps is the most useful for us to focusing on and trying to see how artificial intelligence systems can learn. It's [/INST] interesting. I I do think about this distinction between controlled and automatic or goal directed and habitual behavior a lot in thinking about where we are in A I research. Um but but just to finish finish, the kind of dissertation here, the role of the front of the prefrontal cortex is generally understood these days, sort of in in contradistinction to that habitual domain. In other words, the prefrontal cortex is what helps you override those habits. It's what allows you to say well, what I usually do in this situation is acts, but given the context I probably should do why. I mean the elbow bump is a great example, right? If you know, reaching out and shaking hands is a probably habitual behavior and it's the prefrontal cortex that allows us to bear in mind that there's something unusual going on right now and in this situation I need to not do the usual thing. Uh the kind of behaviors that Lauria reported and he built tests for detecting these kinds of things were exactly like this. So in other words, when I stuck out my hand, I want you instead to present your elbow, a patient with frontal damage would have great deal of trouble with that. You know, somebody proffering their hand would illicit, you know, a handshake. Um the prefrontal cortex is what allows us to say, Hold on hold on, that's the usual thing. But I have the ability to bear in mind even very unusual contexts and to reason about what behavior is appropriate there </s> |
<s>[INST] First is Inside Tracker a service I used to track my biological data. Second is 10,000 the clothes I like to wear for lifting cardio and grappling. Third is for stigmatic, maker of delicious mushroom coffee and plant based protein. Fourth is [/INST] express VPN. </s><s>[INST] The VPN have been using for many years so the choices health, fashion, coffee or digital privacy. She was wisely my friends and now onto the full ad reads as always no ads in the middle. I do try to make these interesting but if you skip them please still check out the sponsors. I enjoy their stuff. Maybe you will too. This show is brought to you by Inside Tracker a service. I used to track bio data. They have a bunch of plans, most of which include a blood test that gives you a lot of information that you can then make decisions based on. They have algorithms, machine learning algorithms that analyze your blood data. DNA data and fitness tracker data to provide you with a clear picture what's going on inside you and to offer you science backed recommendations for positive diet and lifestyle changes. Andrew Huberman the great the powerful Andrew Huberman talks a lot about it. Davidson Claire talks a lot about it. These guys are brilliant. They know what they're talking about, they know what they're doing. So tools like Inside Tracker that allows you to ground your decisions in more and more data that comes from your actual body. The full rich range of biological data that comes from your blood and so on. I think it's obvious that decisions you make about your body should come from that data. I'm a big fan of there's not just for the actual implementation but for the idea itself for a limited time, you can get 25% off the entire inside tracker store. If you go to inside Tracker dot com slash lex. That's inside Tracker dot com slash flex. This show is also brought to you by 10,000 maker of high quality well fitting, comfortable training clothes. I wear their foundation shorts for lifting and cardio and I wear their fight shorts for training jiu jitsu and grappling and actually had somebody a couple of days ago stopped me and recognize the shorts And asked me if I like wearing them like because they heard about 10,000 and they saw a bunch of people wearing them and how do they feel? And I said they feel amazing. I was surprised they recognized it because one of the things I like about 10,000 is um it's very kind of minimalistic, basic and they're hard to see logo to be honest. And the fact that this person recognized the logo and the look of the shorts was pretty cool. It's honestly for both. Like I said, lifting and cardio. Uh and for grappling for jujitsu, all that kind of stuff. It's my favorite. Close to where Anyway, go to 10,000.cc And enter code lex to receive 15% off your purchase. That's 10,000 dot CC and enter code lex. This show is also brought to you by four stigmatic, the maker of delicious mushroom coffee and plant based protein. The coffee doesn't taste like mushrooms in case you were wondering. It tastes delicious. I really enjoyed drinking four stigmatic coffee in the morning is one of the sort of essential rituals in my life as both the warmth, the aroma, the steam coming from the cup of coffee. That's like a Pavlovian signal to my brain that we're going to get some shit done Now, pardon my french, probably the most important moment of my day or stretch of time in my day is a few hours in the morning. Well there's a quiet, deep focus. I try to make that about 34 hour session. If I can get up to 40% off and free shipping on mushroom coffee bundles. If you go to four stigmatic dot com slash flex, that's four stigmatic dot com slash flex. The show is also brought to you by express VPN. I've used them to protect my privacy and the internet for many, many years. It used to be a bright red button. It's more subtle now, but it brought me joy you press the button. That's the only thing it does is uh all of a sudden it makes it appear to the rest of the internet as if you are located somewhere else. It adds that extra layer of protection between you and I. S. P. S. Who want to collect as much data from you as possible even when you're using incognito mode in chrome. Yes that still does not prevent from the sbs collecting your data. Obviously there's also useful aspects to changing your location through the VPN. So like if you're netflix or any kind of other media service you can watch shows that are only available to certain geographical locations. And finally in terms of implementation, I love great design. I love great software and that's what express VPN is. It's fast. It works on any device and operating system including Lennox, my favorite operating system anyway. Good to express VPN dot com slash flex pod to get an extra three months free. That's Express VPN dot com slash flex pod. This is the lex Friedman podcast And here is my conversation with carl hart. I think it is bold and powerful to admit to using in your private life the drugs that you study in your research including heroin and cocaine. So let me ask what is the experience of taking heroin. Like what happens to the body. What happens to the mind when you take it? [/INST] Well you know I take M. D. M. A. Cannabis and all the rest of these drugs to I've tried those drugs um the experience in the body and the mind. I don't really know what people want to know in that regard. It's like saying what is the experience of having an orgasm in the body and the mind or some other sort of event that you really enjoy. Um So I don't really know what </s> |
<s>[INST] Is that what love is too? You were expecting one thing and something else happened. [/INST] I don't know. Yeah. Well, what is it then I think? No, emotion, look, I'm sorry emotion is that but- I think love is just fulfilling your purpose. No, but okay, look, like- Whatever that means that's- I mean really what? So you know- When are you happiest? When you're doing- All right, all right let me go back. Let me go back. So yeah, okay. So, okay, so- Let me go back. Let me define love quickly, okay. Go for it in terms of assembly space. Excellent. Didn't think I'd be doing this today. We... We, man... We can't follow your bliss. Okay! Go for it. That'scheSD. Inte- Okay, okay... I can't wait until Assembly Theory 101 is taught and the second lecture is Assembly Theory of Love. No no! </s><s>[INST] But look, but actually It's being surprised! The expectation's being broken. I'm just.... No, go for it. I'm not.... [/INST] I want to hear you as an emotion. I'm just.. I'm not... No go for it. I want to hear you as an emotionative. I'm not an emotional being. But I would say... so let's talk. So, we'll talk about emotion I had but love is more complex. Love is a very complex set of emotions together and logical stuff. But if you've got this thing, this person that's on this causal chain, that has this empathy for this other thing. Love is being able to project ahead in your assembly space and work out what you're- the person you're in love with has a need for and to do that for them without selflessly, right? Because you can project ahead what they're going to need, and they are there, and maybe you can see someone who's going to fall over and you catch them before they fall off, or maybe you can anticipate that someone's going to be hungry, and without helping you, you just help them. That just sounds like empathy. But it's more complex than that, right? It's more complex. </s><s>[INST] It's more about not just empathy, it's understanding, it's about kind of sharing that experience. [/INST] That just sounds like... Oh, that's not what it's like to feel love, like feeling love is like, like I think it's like when you're aligned with things that you feel like feel love like, are your purpose </s><s>[INST] or your reason for existing? So if you have those feelings towards the robot, why is that rope, I mean, because you said that the AGI will build an AGI, but it won't, there'll be a fundamental difference [/INST] in AGI. I don't think we'll build it. It's going to emerge from our technology. I think you guys will argue the same thing. I just said that GPT, that we do not correctly capture the causal chain that we have. Within GPT. Yeah. </s><s>[INST] Within AI. But don't you think we have. Within GPT. Yeah. I don't think it captures, because GPT-3 is fundamentally trained on a corpus of knowledge, you know, like the internet. Don't you think it gets better and better and better at capturing the memory of all [/INST] the. It will be better at fooling you. And at some point you won't care. But when it comes, my guess, this is a quick, that's all I was getting to right before we got, I got in the love trap. The love trap. Yeah. </s><s>[INST] Like B. Cronin in the love trap, you know, it was like, like B. Cronin in the. [/INST] Sounds like a good, good fan name, a good fan name. Makes it sad. Okay. It's sad. Assembly space is sad. No. It's so short. But I think there are other features that allow, that we pull on innovation, that allow us to do more than what we just see in GPT-3. So if you're being fooled there. So I think what I mean is human beings have this ability to be surprising and creative, whereas is it Dali, this thing, or, or if you take, um, GPT-3 is not going to create a new verb. Shakespeare created new verbs. You're like, wow. And that required Shakespeare to think outside of language in a different domain. So I think having that connections across multiple domains is what you need for AGI. </s><s>[INST] Yeah. But I don't know if you need, um, that I don't know if there's any limitations to GPT, GPT [/INST] 1, not being able to be cost domain. The number one problem is, um, it's instantiated in a resource limited substrate, um, that we don't use in silicon, um, it is, it is the, the architecture is used for training for learning. It is about fooling. It's not about understanding. And I think that there is some understanding that we have that is not yet symbolically </s><s>[INST] representable. But language, learning language and using language seems to be fundamentally about fooling, not understanding. [/INST] What, why, um, why do you use language exactly? I might disagree with that quite fundamentally actually. Um, but I don't, I, I'm not sure I understand how I make a coherent argument for that, but my feeling is that there is, there are, there is comprehension in reality, in our consciousness below language and, and we use those for language, for all sorts of expressions and we don't yet understand that there's a gap. We will get there, but I'm saying, wouldn't it be interesting? It's a bit like saying, could I facsimile you or Sarah into a new human being, right? And let's just say I could copy all your atoms and the positions of all your atoms, electrons into, into this other person. They would be you. The answer's no, and it's quite easy to show using assembly theory, because actually the feature space that you have, that graph, the only way to copy you is to create you on that graph. So everything that's happened to you in your past, we have to have a faithful record for it. One, another copy of Lex. You have to do exact thing. One another copy of Sarah. One, another copy of Lee, the exact past has to be replicated. </s> |
<s>[INST] carried out by the Chinese government, [/INST] this whole thing, which continues to this day, but most intensely in the first months following the outbreak, that's the reason why we have so many dead. And certainly, as I was saying before, I and a small number of others have been carrying this flame since early last year, but it's kind of crazy that our governments haven't been demanding it. And we can talk about the World Health Organization process, which was deeply compromised in the beginning. Now it's become much, much better. But again, it was the pressure of outsiders that played such an important role in shifting our national and international institutions. And while that's better than nothing, it would have been far better if our governments and international organizations </s><s>[INST] had done the right thing from this. If I could just make a couple of comments about Joe Rogan. So there's a bunch of people in my life who have inspired me, who have taught me a lot, who I even look up to. The many of them are alive, most of them are dead. I wanna say that Joe said a few critical words about the conversation with Francis Collins, most of it offline with a lot of great conversations about it, some he said publicly. And he was also critical to say that me asking hard questions in an interview is not my strong suit. And I really want to kind of respond to that, which I did privately as well, but publicly to say that Joe is 100% right on that. But that doesn't mean that always has to be the case. And that is definitely something I wanna work on because most of the conversations I have, I wanna see the beautiful ideas in people's minds. But there's some times where you have to ask the hard questions to bring out the beautiful ideas. And it's hard to do, it's a skill. And Joe is very good at this. He says the way he put it in his criticisms, and he does this in his conversations, which is, whoa, whoa, whoa, stop, stop, stop, stop. There's a kind of sense like, did you just say what you said? Let's make sure we get to the bottom, we'll clarify what you mean. Because sometimes really big negative or difficult ideas can be said as a quick aside in a sentence. Like it's nothing, but it could be everything. And you wanna make sure you catch that and you talk about it. And not as a gotcha, not as a kind of way to destroy another human being, but to reveal something profound. And that's definitely something I wanna work on. I also want to say that, as you said, you disagree with Joe on quite a lot of things. So for a long time, Joe was somebody that I was just a fan of and listened to. He's now a good friend. And I would say we disagree more than we agree. And I love doing that. And, but at the same time, I learned from that. So it's like dual, like nobody in this world can tell me what to think. But I think everybody has a lesson to teach me. I think that's a good way to approach it. Like whenever somebody has words of criticism, I assume they're right and walk around with that idea to really sort of empathize with that idea because there's a lesson there. And oftentimes my understanding of a topic becomes, is altered completely or it becomes much more nuanced or much richer for that kind of empathetic process. But definitely, I do not allow anybody to tell me what to think, whether it's Joe Rogan or Fyodor Dostoyevsky or Nietzsche or my parents or the proverbial girlfriend of which I don't actually have. [/INST] But she's still busting my balls. Exactly, exactly. </s> |
<s>[INST] suffering and the fixes in fixing the habits just uh [/INST] I'm still working on the fix first of all. So there's okay. There is um there's just that you have a boat that sucks and it's leaking and it's gonna sink. You can fix it with duct tape for a couple of you know for one ride or whatever that's not really fixing the boat but you can can get you by. So there's duct tape solutions to me. So the panic monster is the character that rushes into the room once the deadline gets too close or once there's some scary external pressure not just from yourself and that's a huge aid to a lot of procrastinators. Um Again there's a lot of people who won't you know do that thing they've been writing that book they wanted to write but there's way fewer people who will not show up to the exam. You know most people show up to the exam. So um that's because the panic monster is gonna freak out if they don't. So you can then you can create a panic monster if you wanna, you know, you really want to write music, you really want to become a singer songwriter. Well, book a venue, tell 40 people about it and say, hey, you know, this day, two months from now come, come and see, I'm gonna play you some of my songs, you know, have opinions, you're gonna write songs, you're gonna have to write. So, um, there's there's there's duct tape things, you know, you can do things, you know, people do. There's, I've done a lot of this with a friend and I say if I don't get X done by a week from now, I have to donate a lot of money somewhere I don't want to donate. And </s><s>[INST] that's you put them in the category of duct tape solutions. [/INST] Yeah. Because it because it's not why do I need that, right? If I really assault this, this is something I want to do for me. It's selfish. This is I just literally just want to be selfish here and do the work I need to do to get the goals I want to get right there. There's there's a much, all the incentives should be in the right place. And yet if I don't say that I will, it'll be a week from now and I won't have done it. Something weird is going on. There's some resistance, there's some force that is prevent that is in my own way, right? And so doing something where I have to pay all this money. Okay now I'll panic and I'll do it. So that's duct tape fixing their boat. Is something where I don't have to do that. I just will do the things that I again, it's not, I'm talking about super crazy work ethic. Just like for example, okay, I have a lot of examples because I have a serious problem that I've been working on. Um and in some ways I've gotten really successful at solving it in other ways. I'm still still floundering. So you're the </s><s>[INST] world's greatest duct taper. [/INST] Yes. Well I'm pretty good at duct taping. Probably could be even better. And I'm like, and I'm and I'm you're procrastinating </s><s>[INST] and becoming a better [/INST] duct tape literally like yes, there's nothing I won't. So here's what I know what I should do as a writer, right? It's very obvious to me is that I should wake up. I don't have to be crazy about six a.m. or anything saying or I'm not going to be one of those Crazy people of 5:30 Jogs. I'm gonna wake up at whatever, you know, 7 38 8 30 And I should have a block like just say 9 to noon where I get up and I just really quick make some coffee and and right, </s><s>[INST] it's obvious because all the great writers in history did exactly that some [/INST] some of them have done that, that that's common, there's there's something that I like these writers, they do the late night sessions, but most of them they do, </s><s>[INST] there's a session, but there's a session that's [/INST] most writers write in the morning and there's a reason I I don't think I'm different than those people. Um it's a great time to write your fresh, write your ideas from the from dreaming have kinda collected you have all, you know, new answers that you didn't have yesterday and you can just go. But more importantly, if I just had a routine where I wrote from 9 to noon weekdays, Every week would have a minimum of 15 focused hours of writing, which doesn't sound like a lot, but it's a lot of 15, 15. No, this is no joke, this is, you know, you're not your phones away, you're not talking to anyone, you're not opening your email, you are focused writing for three hours five. That's a big week for most writers. Right? So now what's happening is that every weekday is a minimum of a B. I'll give myself, you know, what a might be, you know, wow, I really just got into a flow and run for six hours and had great, but it's a minimum of a B. I can keep going if I want and every week is a minimum of a B without 15 hours, right? And if I just had to talk about compiling, if this is the two pages a week, if I just did that every week I achieved all my writing goals in my life. And yet I wake up and most days I just either I'll revenge procrastination later and I can go to bed way too late and then wake up later and get on a bad schedule and I just fall into these bad schedules or I'll wake up and there's just, you know, I'll say I was gonna do a few emails and I'll open it up and I have something on texting and texting and or I'll just go and you know, I'll make a phone call and I'll be on phone calls for three hours is always something or I'll start writing. And then I hit a little bit of a wall. But because there's no sacred, this is a sacred writing block, I'll just hit the wall and say, well this is icky. And I'll go do something else. So duct tape what I've done is um, uh they're white, but white has one employee Alicia, she's the manager of lots of things. That's her role. She truly does lots of things. Um And one of the things we started doing is Either she comes over and sits next to me where she can see my screen from 9 to PM. That's all it takes. The thing about procrastination is, there's usually they're not kicking and screaming, I don't want to do this, It's the feeling of, you know, in the old days when you had to go to class, you know, your lunch block is over and it's like ship class in five minutes or it's monday morning. And uh but you said you know what, but you know you go, you say okay and then you get to class and it's not that bad once you're there, right, you know, you have a trainer and he said okay next set and you go okay and you do, that's all it is. It's someone some external thing being like, okay, I have to do this. And then you have that moment of like sucks, but I guess I'll do it if no one's there though. The problem with a procrastinator is they don't have that person in their head. Other people I think were raised in the sense of shame if they don't do stuff and that stick in their head is hugely helpful. I don't really have that. Um And so anyway, Alicia is sitting there next to me, it's not, she's she's doing her homework but she can see my screen and she of all people knows exactly what I should be doing when I shouldn't be doing. That's all it takes the shame of just having her see me while she's sitting there not working would just be too, it's too weird and too embarrassing. So I get it done and it's amazing. It's like a game changer for me. So duct tape can solve sometimes duct tape is enough. But I'm curious to, I'm still trying to what is going on. I think part of it is that we are actually wired. I think I'm being I'm being very sane human actually is what's happening. You're not saying it's not the right word. I'm being like, and being a natural human that we are not programmed to sit there and do homework of a certain kind that we get the results like six months later, like that is not so we're supposed to, you know, conserve energy and like fulfill our needs as we need them and like do immediate things and um we're overriding our natural ways when we wake up and get to it. And I think sometimes because the pain, I think a lot of times we're just avoiding suffering. And a lot of a lot of people, the pain of not doing it is actually worse because they feel shame. So if they don't get up and take a jog and get up early and get to work, I will feel like a bad person and that is worse than doing those things. And then it becomes a habit, eventually it becomes just easy, automatic becomes I do it because that's what I do. But I think that if you don't have a lot of shame necessarily the pain of doing those things is worse in that in the immediate moment than not doing it and </s> |
<s>[INST] Yeah, I find that whole project of the human species quite inspiring these like huge Moonshot projects. Somebody I was reading something um, in terms of the source of food that that may be the most effective on mars is you could farm insects. That's the easiest thing to farm. So we'll be eating like cockroaches before living on MArs because that's the easiest thing to actually um, as a source of protein. So growing a source of protein is the easiest thing is insects. I just imagine this giant for people who are afraid of insects, this is not a place, maybe you're not supposed to even think of it, that will be like a cockroach, milkshake or something like that, [/INST] right? I wonder if they have people been working on the genetic engineering of of insects to make them </s><s>[INST] radiation from [/INST] uh resistant or whatever it </s><s>[INST] is. What can possibly go wrong speaking, radiation resistant, they're already like, survive everything. Plus I uh I took an allergy test in Austin. So there's everybody's is like the allergy levels are super high there. Uh And uh one of the things, apparently I'm not allergic to any insects except cockroaches. Hilarious. So maybe uh um well, I'm going to use that as you know, people use uh an excuse that I'm allergic to cats to not have cats. I'm gonna use that as an excuse to uh not go to mars as one of the first batch of people. I was [/INST] gonna ask if you had the opportunity when you go. </s><s>[INST] Yeah, I'm joking about the cockroach thing and I would definitely go. I love challenges. I love I love things. I love doing things where the possibility of death is is uh not insignificant because it makes me appreciate it more det, meditating on death, makes me appreciate life. And when the meditation on death is forced on you, because of how difficult the task is. I enjoy those kinds of things. Most people don't, it seems like, but I love the idea of difficult journeys. Um for no purpose whatsoever except exploration going into the unknown, seeing what the limits of the human mind and the human body are. It's like what the hell else is this whole journey that we're on for like uh but it could be because I grew up in the soviet Union, there is a kind of love for space, like the space race, the Cold war created. I don't know if still it permeates american culture as much. Oh, especially with the data as a scientist. I think I've I've loved the idea of humans driving out towards the stars. Always like from the engineering perspective has been really exciting. I don't know if people love that as much in America anymore, I think Ellen is bringing that back a little bit, that excitement about rockets and going out there. But uh so that's that's helpful for me. I always loved that idea from alien scientist perspective. If you were to look back on earth, is there something interesting you could say about earth? Like how would you summarize earth? Like, you know, like Hitchhiker's guide to the galaxy, like if you had to report, Like write a paper on earth or like a letter, like a like a one pager um summarizing the contents of the surface and the atmosphere. Is there is there something interesting like, do you ever take that kind of perspective on it? I know you like volcanism, so volcanoes that will probably be in the report. [/INST] I was going to say that's where I was going to go first. There a few things to say about the atmosphere, but in terms of the volcano. So one of the really interesting puzzles to me and planetary sciences. So we can, we can look out there and we've been talking about surfaces and volcanoes and atmospheres and things like that. But that is just this tiny little veneer on the outside of the planet and most of the planet is completely inaccessible to telescopes or two spacecraft missions. You can drill a meter into the surface, but that's still really the veneer. Um, and one of the cool puzzles is looking at what's going on on the surface and trying to figure out what's happening underneath. Or just any kind of indirect means that you have to study the interior because you can't dig into it directly, even on Earth, you can't dig deep into Earth. So, from that perspective, looking at Earth. Um, one thing that you would be able to tell from orbit given enough time is that Earth has tectonic plates. So you would see that volcanoes follow the edges if you trace for all the volcanoes on earth, they follow these lines that trace the edges of the plates. And similarly, you would see things like the Hawaiian string of volcanoes that you could in for just like, you know, we did as people actually living on Earth, that the plates are moving over some plume that's coming up through the mantle. And so you could use that to say if the aliens could look at where the volcanoes are, are happening on earth and say something about the fact that Earth has plate tectonics, which makes it really unique in the solar system. </s><s>[INST] So the other plants don't have, [/INST] it's the only one that has plate tectonics. Yeah. </s><s>[INST] What about IO and the friction and all that? That's not what's the difference between. So all this plate tectonics, like another layer of like solid rock that moves around and there's cracks. [/INST] Yeah. So, so earth has plates of solid rock sitting on top of a partially molten layer. And those plates are kind of shifting around um on IO. It doesn't have that. And the volcanism is what we call heat pipe. Volcanism, it's the magma just punches a hole through the crust and comes out on the surface. I mean, that's a simplification, but that's effectively what's happening </s> |
<s>[INST] a million per ton. [/INST] Yes, ideally less. Much less than a million tons. But if it's not like it's going to be well how much can society afford to spend or just want to spend on a self sustaining city on mars, the self sustaining part is important, like it's just the key threshold. Um The grateful to will will have been passed when the city on mars it can survive even if the spaceships from a stop coming for any reason, it doesn't matter what the reason is, but if they stop coming for any reason will it die out or will it not? And if there's even one critical ingredient missing, then it still doesn't count. It's like, you know if you're on a long sea voyage and you've got everything except vitamin C. It's only a matter of time, you know you're gonna die. So so we're gonna get Morris Amauri city to the point where it's self sustaining um I'm not sure this will really happen in my lifetime, but I hope to see it at least have a lot of momentum and and then you can say, okay, what is the minimum tonnage necessary to have a self sustaining city? Um and there's a lot of uncertainty about this. You can, I don't know, it's probably at least a million tons because you have to set up a lot of infrastructure on on MARS. Um like I said, you can't be missing any anything that in order to be self sustaining, you can't be like you need a semiconductor fabs, you need iron ore refineries, like you need lots of things, you know, So um and MArs is not super hospitable. It's it's the least inhospitable planet, but it's definitely a fixer upper of a planet </s><s>[INST] outside of Earth. [/INST] Yes. Earth is pretty like easy. And </s><s>[INST] also should, we should clarify in the solar system, Yes, in the solar system, there might be nice like vacation spots, [/INST] there might be some great planets out there, but it's hopeless </s><s>[INST] to get their way [/INST] way, way, way, way too hard to say the least. </s><s>[INST] Let me push back on that. Not really a pushback but a quick curveball of a question. So you didn't mention physics as the first starting point. So general relativity allows for warm holes. They technically can exist. Do you think? Um those can ever be leveraged by humans to travel faster than the speed of light. [/INST] Well </s><s>[INST] the whole thing [/INST] is, is debatable. Uh the, that we currently do not know of any means of going faster than the speed of light there, there is like, like there are some ideas about having space like so, so you're gonna move at the speed of light through through space, but if you can make space itself move that, that that's like that, that's what warping space um spaces is capable of moving faster than the speed of light. Like the universe in the big Bang, the universe expanded it much, much more than the speed of light by a lot. Yeah. Um so, but the, If this is possible that the amount of energy required 12 space is so gigantic, it's boggles the mind. </s><s>[INST] So all the work you've done with propulsion, how much innovation is possible with rocket propulsion is this? Um I mean you've seen it all and you're constantly innovating in every aspect. How much is possible? Like how much can you get 10 x somehow. Is there something in there in physics that you can get significant improvement in terms of efficiency of engines and all those kinds of things? [/INST] Well, as I'm saying like the really the Holy Grail is a fully and rapidly reusable orbital system. Um so, uh right now The Falcon nine is the only reusable rocket out there, but the booster comes back and lands and you've seen the videos uh and we get the nose cone or fairing back but we do not get the upper stage back. So ah, that means that we have a minimum cost of building up for stage. Um You can think of like a two stage rocket of sort of like to our planes, like a big airplane and a small airplane. Um, and we get a big airplane back, but not the small airplane and so it still costs a lot, you know, so that upper stages, You know, at least $10 million. Um, and then the degree of the booster is not as reused, it's not as rapidly and completely reusable as we'd like and or the Ferencz, so, you know, are kind of minimum marginal cost, not counting overhead for perf lighters on the order of $15 - $20 million dollars maybe. Um, So, uh, that's, that's extremely good for, it's by far better than any rocket ever in history. Um, but uh, with full and rapid reusability, we can reduce the cost per ton to orbit by A factor of 100. Just think of it like um, like imagine if you had an aircraft or something or a car. Oh yeah, and if you had to buying your car every time you went for a drive would be very expensive. Very silly frankly. But, but in fact you just refuel the car or recharge the car and that makes your trip like 1000 times cheaper. So it's the same for rockets if you, it's very difficult to make this complex machine that can go to orbit and so if you cannot reuse it and have to throw even any part of any significant part of it away that massively increases the cost. So, you know, starship in theory could do a cost per launch of like a million, maybe $2 million or something like that and and put over 100 tons in orbit, which is crazy. </s> |
<s>[INST] That was a beautifully articulated vision for the world. So thank you for that. [/INST] Thank you for that. Let's talk a little bit about life and death. I'm pro-life and anti-death. For most people, there's few exceptions and I won't mention here. </s><s>[INST] I'm glad just like your dad, you're taking a stand against death. By the way, you have a bunch of awesome music where you play piano online. Well, one of the songs that I believe you've written, the lyrics go, by the way, I like the way it sounds. People should listen to it. It's awesome. I was, I considered, I probably will cover it. It's a good song. Tell me why do you think it is a good thing that we all get old and die? It's one of the songs. I love the way it sounds, but let me ask you about death first. Do you think there's an element to death that's essential to give our life meaning? The fact that [/INST] this thing ends? Let me say, I'm pleased and a little embarrassed you've been listening to that music I put online. That's awesome. One of my regrets in life recently is I would love to get time to really produce music well. I haven't touched my sequencer software in like five years. I would love to rehearse and produce and edit, but with a two year old baby and trying to create the singularity, there's no time. So I just made the decision to, when I'm playing random shit in an off moment. Just record it. Just record it. Whatever. Maybe if I'm unfortunate enough to die, maybe that can be input to the AGI when it tries to make an accurate mind upload of me. Death is bad. I mean, that's very simple. It's baffling. We should have to say that. I mean, of course people can make meaning out of death. And if someone is tortured, maybe they can make beautiful meaning out of that torture and write a beautiful poem about what it was like to be tortured, right? I mean, we're very creative. We can milk beauty and positivity out of even the most horrible and shitty things. But just because if I was tortured, I could write a good song about what it was like to be tortured. It doesn't make torture good. And just because people are able to derive meaning and value from death doesn't mean they wouldn't derive even better meaning and value </s><s>[INST] from ongoing life without death. Which I very... So if you could live forever, that's awesome. It was like a death is bad. I mean, that's very which I very definite. Yeah. Yeah. So if you could [/INST] live forever, would you live forever? Forever. I, my, my goal with longevity research is to abolish the plague of involuntary death. I don't think people should die unless they choose to die. If I had to choose forced immortality versus dying, I would choose forced immortality. On the other hand, if I chose, if I had the choice of immortality with the choice of suicide, whenever I felt like it, of course, I would take that instead. And that's the more realistic choice. I mean, there's no reason you should have forced immortality. You should be able to live until you get sick of living, right? And that will seem insanely obvious to everyone, 50 years from now, and they will be so... I mean, people who thought death gives meaning to life, so we should all die, they will look at that 50 years from now, the way we now look at the Anabaptists in the year 1000, who gave away all their positions, went on top of the mountain for Jesus to come and bring them to the ascension. I mean, it's ridiculous that people think death is good because you gain more wisdom as you approach dying. I mean, of course it's true. I mean, I'm 53, and the fact that I might have only a few more decades left, it does make me reflect on things differently. It does give me a deeper understanding of many things, but I mean, so what? You could get a deep understanding in a lot of different ways. Pain is the same way. We're going to abolish pain, and that's even more amazing than abolishing death, right? I mean, once we get a little better at neuroscience, we'll be able to go in and adjust the brain so that pain doesn't hurt anymore, right? And that, you know, people will say that's bad because there's so much beauty in overcoming pain and suffering. Oh, sure, and there's beauty in overcoming torture too, and some people like to cut themselves, but not many, </s><s>[INST] right? I mean, to the ascension. That's an interesting, but to push back again, this is the Russian side of me. I do romanticize suffering. It's not obvious. I mean, the way you put it, it seems very logical. It's almost absurd to romanticize suffering or pain or death, but to me, a world without suffering, without pain, without death, it's not obvious what that [/INST] you can say in the people's Zoo. The people torturing each other. </s><s>[INST] I guess what I'm trying to say, I don't know if I was presented with that choice what I would choose [/INST] because to me, this is a subtler matter, and And I've posed it in this conversation in an unnecessarily extreme way. So I think the way you should think about it is what if there's a little dial on the side of your head and you could turn to how much pain hurt? Turn it down to zero, turn up to 11, like in spinal tap if it wants, maybe through an actual spinal tap, right? So I mean, would you opt to have that dial there or not? That's the question. The question isn't whether you would turn the pain down to zero all the time. Would you opt to have the dial or not? My guess is that in some dark moment of your life, you would choose to have the dial implanted </s> |
<s>[INST] Yeah, because I from a psychological perspective, it's just gonna be a stressful mess as people have studied, right? It's like if people are most likely the process of colonization, like looks like basically a prison, like you're in a very tight enclosed space with people and it's just a really stressful environment, you know, how do you select the kind of people that will go and then there will be drama. There's always drama and I just want to show up and there's some rules but you know, it depends. So I'm not worried about the health and the technical difficulties. I'm more worried about the psychological difficulties and also just not being able to tweet like what are you gonna how are you talking? There's no netflix. So yeah, maybe not in the first million, but the 1st 100,000. It's exciting to define the direction of a new, like how often do we not just have a revolution to redefine our government as you know, smaller countries are still doing to this day literally start over from scratch. There's uh just our financial system, it could be like based on the Cryptocurrency, you can think about like how democracy, you know, we have, we have now the technology that can enable pure democracy for example, if we choose to do that as opposed to representative democracy, all those kinds of things. So we talked about two different forms of propulsion which you're super exciting. So the chemical based that's doing pretty well and then the electric bass is um are there types of propulsion that might sound like science fiction right now but are actually within the reach of science to the next 10 2030, 50 years that you kind of think about or maybe even within the space of even just like like even ion engines, Is there like breakthroughs that might 10 X the thing like really improve it. [/INST] So you know, the real game changer would be propellant liss propulsion and so every couple of years you see a new now a startup or a researcher comes up with some contraption for producing thrust that didn't require. You know, we've been talking about conservation of momentum mass, times velocity out the back. Um Masters usually forward. </s><s>[INST] Yes [/INST] exactly and you have to you know carry that up with you or find it on an asteroid or harvest it from somewhere if you didn't bring it with you. So not having to do that would be you know, one of the ultimate game changers. Um and and I you know, unless there are new types of physics, I don't know how we do it but it comes up often. So it's something I do think about and you know the one I think it's called the casimir effect. If you can if you have two plates and the space between them is on the order of these like the wavelength of these ephemeral vacuum particles that pop into and out of existence or something. Um I may be confusing multiple types of propellant lys forces but um that that could be real and could be something that that we use eventually </s><s>[INST] will be the power [/INST] source. Yeah. The most recent engine like this that has was just debunked this year I think in in March or something was called the em drive and um supposedly you used a power source so you know, batteries or solar panels to generate microwaves into this resonant cavity. And people claimed it produced thrust. So they went straight from this really loose concept to building a device and testing it and they said we've measured thrust and sure on their thrust balance they saw thrust and different researchers built it and tested it and got the same measurements and so it was looking actually pretty good. No one could explain how it worked. But what they said was that um this inside the cavity um The microwaves themselves didn't change but the speed of light changed inside the cavity so relative to that um You know their momentum was conserved. Um And I don't know whatever. Um But finally someone I think at Nasa built the device, tested it, got the same thrust, then unhooked it flipped it backwards and turned it on but got the same thrust in the same direction again. And so they're like this is just an interaction with the test set up or you know some of the chamber or something like that so um forwarded again. Um But you know it would it would be so wonderful for everybody if we could figure out how to do it. But I don't know. </s><s>[INST] That's an interesting twist on it because that's more about efficient travel. Long distance travel. Right? That's not necessarily about speed. That's more about enabling like let's [/INST] hook that up to the nuclear power supply. There you go. </s><s>[INST] Okay. But still in terms of speed in terms of trying to. So there's recently I was already I think been debunked or close to being the bunk. But the signal, a weird signal from our nearby friends, nearby exoplanets from Proxima centauri ah A signal that's 4.2 light years away. So, you know, the thought is uh it'd be kind of cool if there's life out there, alien life, but would be really cool if you could fly out there and check and so what kind of propulsion? And do you think about what kind of propulsion allows us to travel close to the speed of light or, you know, half the speed of light, all those kinds of things that would allow us to get to Proxima centauri and unreasonable in a lifetime. [/INST] You know, there's the project breakthrough star shot um that's looking at sending those tiny little chip sets um Accelerating really fast. Yeah. Using a laser, so launching them and then while they're still relatively close to the earth blasting them with some, I forget what even, what power Level you needed to accelerate them fast enough to get there in 20 years. Crazy </s> |
<s>[INST] What do you think is? Uh, we've been talking about how to extend this life, how to make it a good life. I've been talking about uh, love, what do you think is the meaning of this whole ride of life of life? Why do we want to make it a good one? Do you think there's a meeting, Do you think there's an answer to the why? [/INST] Okay, for me personally, I think the, the meaning of life is, is to eat, is to sleep, is to fall in love, is to cry and then to die? Oh, and probably race Cause in between </s><s>[INST] these cars. Well, there's a whole topic of sex we didn't talk about probably. [/INST] So should we </s><s>[INST] do that? Well, maybe if you'll have me back that I will give around next time. We will do another three hours on Sex alone. Yeah, it has been uh over three hours. Gosh, I'm a big fan of your work. I think you're doing really important work, even despite all the things I've been saying about the madness of my own sleep schedule, I think you're helping millions of people. So it's an honor that you spend your valuable time with me and I can't wait until your podcast comes out. I'm a huge fan of podcast. I'm a huge fan of you and it's just an honor to know you and to get a chance. Hopefully in the future to work together with you. You're [/INST] a brilliant man and you're doing amazing things and I feel immensely on it to have I met you too. Now know you and um, to start calling you a friend. Thank you for what you do for the world. And um, for me included, </s><s>[INST] thank you, Matt, take care. Thanks for listening to this conversation with matt walker and thank you to stamps dot com squarespace letter Green's better help. And on it, check them out in the description to support this podcast. And now let me leave you with some words from Nicola Tesla who we discussed in this podcast as sleeping very few hours a night. All those great in the past was ridiculed, condemned, combated and suppressed only to emerge all the more powerfully, all the more triumphantly from the struggle. Thank you for listening and hope to see you next time. [/INST] </s> |
<s>[INST] passing through. So after you do the clang parsing, what's what's the graph? What does it look like? What's the data structure [/INST] here? Yeah. So in in the purser it's usually a tree, it's called an abstract syntax tree. And so the idea is you you have a node for the plus that the human wrote in their code or the function call, you'll have a note for a call with the function that they call and the arguments they pass, things like that. This then gets lowered into what's called an intermediate representation. And intermediate representations are like, Alabama has one and there it's a it's what's called a control flow graph. And so you represent each operation in the program as a very simple, like this is going to add two numbers. This is going to multiply two things. Maybe we'll do a call but then they get put in what are called blocks. And so you get blocks of these straight line operations or instead of being nested like in a tree, it's straight line operations. And so there's a sequence in ordering to these operations. </s><s>[INST] And then in the block we're outside the block [/INST] that's within the bloc. And so it's a straight line sequence of operations within the block. And then you have branches like conditional branches between blocks. And so when you write a loop, for example, um and the syntax tree, you would have a four node like for four statement and a c like language, you have a four node and you have a pointer to the expression for the initialize er a pointer to the expression for the increment, a pointer to the expression for the comparison. A pointer to the body. Okay. And these are all nested underneath it in a control flow graph, you get a block for the code that runs before the loop. So the initialize our code, then you have a block for the body of the loop and so the body of the loop code goes in there, but also the increment and other things like that. And then you have a branch that goes back to the top and a comparison and branch that goes out and so it's more of an assembly level kind of representation. But the nice thing about this level of representation is it's much more language independent. And so there's lots of different kinds of languages with different kinds of um you know, javascript has a lot of different ideas of what is false, for example and all that can stand the front end but then that middle part can be shared across all </s><s>[INST] those. How close is that intermediate representation to uh two Neural networks for example, is there are they? Because everything you described as a kind of a at close of neural network graph? I think we are their neighbors or [/INST] what they're quite different in details but they're very similar and idea. So one of the things that neural networks do is they learn representations for data at different levels of abstraction. Right. And then they transform those through layers. Right. Um so the compiler does very similar things. But one of the things the compiler does is it has relatively few different representations or a neural network. Often as you get deeper, for example, you get many different representations in each, you know, layer or set of ops is transforming between these different representations and compiler often you get one representation and they do many transformations to it. And these transformations are often applied iterative lee um and for programmers, they're familiar types of things, for example, trying to find expressions inside of a loop and pulling them out of the loop so they execute for your times or find redundant computation or find um uh constant folding or other simplifications turning, you know, two times X into X, shift left by one. And and things like this are all all all the examples of the things that happen. But compilers end up getting a lot of fear improving and other kinds of algorithms that try to find higher level properties of the program that then can be used by the optimizer. </s><s>[INST] Cool. So what's like the biggest bang for the buck with optimization? What's [/INST] uh </s><s>[INST] Well, no, not even today, at the very beginning of the 80s, I don't know. But [/INST] Yeah, so for the 80s, a lot of it was things like register allocation. So the idea of in in a modern, like a microprocessor, what you'll end up having is you're having memory, which is relatively slow and then you have registers relatively fast, but registers you don't have very many of them. Okay. And so when you're writing a bunch of code, you're just saying, like compute this, put in temporary variable computers, computers compute this. Put in temporary variable. I have a loop, I have some other stuff going on. Well now you're running on an X- 86, like a desktop PC or something? Well, it only has in some cases some modes, eight registers. Right. And so now the compiler has to choose what values get put in what registers at what points in the program. And this is actually a really big deal. So if you think about you have a loop and then an inner loop that executes millions of times, Maybe if you're doing loads and stores inside that loop, then it's gonna be really slow. But if you can somehow fit all the values inside that loop. And registers now it's really fast. And so, getting that right requires a lot of work because there's many different ways to do that and often what the compiler ends up doing is it ends up thinking about things in a different representation than what the human wrote right, you wrote into X. Well, the compiler thinks about that is four different values, each which have different lifetimes across the function that it's in and each of those could be put in a register or memory or different memory or maybe in some parts of the code re computed instead of stored and reloaded. And there are many of these different kinds of techniques that can be used, </s> |
<s>[INST] Yeah, it does seem to me that exomoons open wider or open for the first time, the door to aliens. So more seriously, academically studying, all right, let's look at alien worlds. So I think it's still pretty fringe to talk about alien life, even on Mars and the moons and so on. You're kind of like, you know, it'll be nice. But imagine the first time discover a living organism that's going to change, then everybody will look like an idiot for not focusing everything on this. Because the possibility of the things will, it's possible it might be super boring. It might be very boring bacteria, but even the existence of life elsewhere, I mean, that changes everything. That means life is everywhere. [/INST] Yeah, if you knew now that in five years, ten years, the first life would be discovered elsewhere, you knew that in advance, it would surely affect the way you approach your entire career, especially someone junior in astronomy, you would surely be like, well, this is clearly going to be the direction I have to dedicate my classes and my training and my education </s><s>[INST] towards that direction. All the new textbooks, all the written. And I think there's a lot of value to hedging, like allocating some of the time to that possibility, because the kind of discovery will, the kind of discoveries we might get in the next few decades, it feels what like we're on the verge of a lot of getting a lot of really good data and having better and better tools that can process that data. So there's just going to be a continuous increase of the kind of discoveries that will open, but a slam dunk, that's hard to come by. [/INST] Yeah, I think a lot of us are anticipating, I mean, we're already seeing it to some degree with Venus and the phosphine incident, but we've seen it before with Bill Clinton and the White House lawn announcing life from Mars. And there are inevitably going to be spurious claims or at least claims which are ambiguous to some degree. There will be for sure a high profile journal, like Nature or Science, that will one day publish a paper saying, buy a signature discovered or something like that on Trappist-1 or some other planet. And then there will be years of back and forth in the literature. And that might seem frustrating, but that's how science works. That's the mechanism of science at play, of people scrutinizing the results to intense skepticism. And it's like a crucible, you burn away all the relevances until whatever is left is the truth. And so you're left with this product, which is that, okay, we either believe or don't believe that biote signatures are there. So there's inevitably going to be a lot of controversy and debate and argument about it. We just have to anticipate that. And so I think you have to basically have a thick skin to some degree academically to dive into that world. And you're seeing that with phosphine. It's been uncomfortable to watch from the outside the kind of dialogue that some of the scientists have been having with each other about that because- They get a little aggressive. Yeah. And you can understand why because- </s><s>[INST] Jealousy? I don't know. That's me saying, not you. That's me talking. [/INST] I'm sure there's some envy and jealousy involved on the behalf of those who are not part of the original discovery. But there's also, in any case, just leave the particular people involved in Venus alone. In any case of making a claim of that magnitude, especially life, because life is pretty much one of the biggest discoveries of all time. You can imagine scientifically. You can see, and I'm so conscious of this in myself when I get close to, as I said, even the much smaller goal of setting an exomoon, the ego creep in. And so as a scientist, we have to be so guarded against our own egos. You see the lights in your eyes of a Nobel Prize or the fame and fortune and being remembered in the history books. And we all grew up in our training learning about Newton and Einstein, these giants of the field, Feynman and Maxwell. You get the idea of these individual contributions which get immortalized for all time, and that's seductive. It's why many of us with a skill set to go into maybe banking instead decided, actually, there's something about the idea of being immortalized and contributing towards society in a permanent way that is more attractive than the financial reward of applying my skills elsewhere. So to some degree, that ego can be a benefit because it brings in skillful people into our field who might otherwise be tempted by money elsewhere. But on the other hand, the closer you get towards when you start flirting with that Nobel Prize in your eyes or you think you're on the verge of seeing something, you can lose objectivity. A very famous example of this is Barnard's Star. There was a planet claimed there by Peter van der Kampf, I think it was in 1968, 69. And at the time, it would have been the first ever exoplanet ever claimed. And he felt assured that this planet was there. He was actually using the wobbling star method, but using the positions of the stars to see them to claim this exoplanet. It turned out that this planet was not there, there. Subsequent analyses by both dynamicists and theorists and those looking at the instrumental data established fairly unanimously that there was no way this planet was really there. But Peter van der Kampf insisted it was there despite overwhelming evidence that was accruing against him. And even to the day he died, which was, I think in the early 90s, he was still insisting this planet was there, even when we were starting to make the first genuine exoplanet discoveries. And even at that point, I think Hubble had even looked at that star and had totally ruled out any possibility of what he was talking about. And so that's a problem. How do you get to a point as a scientist where you just can't accept anything that comes otherwise, because it starts out with the dream of fame, and then it ends in a stubborn refusal to ever back down. </s> |
<s>[INST] You showed up in a robe, which I thought was kind of weird. And you had a crown and you kept calling yourself king. [/INST] Correct, and that was in a robe that was a yukata. Okay, thank you for clarifying. Wait, audience, he's joking. Read it. </s><s>[INST] Okay, thank you for clarifying. I heard somebody's gonna write blog posts. Charles Hoskinson walks around with a crown. Let me ask you a ridiculously big and the most important question, what's the meaning of this whole thing? Of life. Well, I got a story. [/INST] Well, I got a story for that. </s><s>[INST] Does this have something to do with the farm? [/INST] Well, no, no, no, it's from Japan. I used to live in Japan. I lived in Osaka, next to Namba, a beautiful area. It's like, if you're gonna live anywhere in Japan, live in Osaka. You live anywhere in Osaka, live next to the restaurant district. It's four o'clock in the morning and get good ramen. These are the things in life that make you who you are. So there was a shogun, and he was kind of a badass. He was really good at killing people, really good at running his empire. And then he got a bit disgruntled in his late forties and he said, you know, I'm just gonna give it all to my son and I'm gonna wander around Japan until I find the perfect cherry blossom. That's what I'm going to do. Everybody thought he went crazy. He said, No, no, no, that's what I'm going to do. And he took his whole entourage with him. And so his son is now the shogun and then he's just wandering throughout Japan and having all these incredible, crazy adventures as he's wandering throughout Japan. Fighting bandits and loving beautiful women and so forth. And then 30 years later, he's passing these two geisha gals and one of them turns around and they notice him slumped over next to a cherry tree. And so she goes over to try to rouse him and he's dead. And in his hands is a wilted cherry blossom. And in his hands, it's a very Japanese story, right? The point is that it's not the actual blossom finding the perfection that matters. It's the things you do on a day-to-day basis. The places you go, the people you meet, the experiences you have, and the joy you take in the things that you do here in the moment now. You have to get there. You know, if you look at Jiro Dreams of Sushi, there's this guy 70 years of his life making the same damn piece of sushi again and again and again. He's the happiest guy around. Albert Camus and his story of Sisyphus. You know, Sisyphus should be miserable. It's the Greek curse. No, he's happy. He has total clarity of purpose. And every day he gets to basically roll that stone just a little bit better than the day before, a different way than the day before. And it's not the destination. It isn't getting the stone up to the top of the hill that matters. It's the fact that the act, you find that joy and that act, that ikigai, that way of life, that purpose of life. That I think is the closest thing a human can get to a meaning. You're a blip. You know, you didn't exist. You're dead. And if you compare it to the size of the universe in that time, it's just a little blip. So all you can do with what you have is just find meaning in the things that you do on a daily basis. And you can't predict the macro. We have all this wealth and power in America. What if a world war breaks out? We could be destitute like Walmart, Germany. And so, you know, you could be a big guy. You could be now living on the street side. Would you be miserable? The point of life is getting to a point that no matter what comes your way, what misfortune comes your way, you're in a position where you can find a modicum of happiness and love and empathy for others in that moment. And then the highest pursuit of life is the ability to share that mindset with other people and give it to them somehow. And that's really hard, you know, because everybody comes in, they're always eeyore. Oh, doom and gloom and cynical. Yeah, but this and that and reputation this and that. You have to somehow transcend all of it and say, you know, that doesn't matter. Look at the cherry blossoms. Aren't they beautiful today? And let's find a better one tomorrow. </s><s>[INST] 70 years of his life. See, you're also a fan of fishing, I've read somewhere. And one of my favorite books is Old Man in the Sea, where there's an old man sort of battling big fish and basically closing out the last chapter of his life in this battle. So I think another aspect of life with this boulder, it feels like the boulder gets bigger and bigger as we get closer to death. And, you know, you find yourself married to a particular struggle in life that eventually just kind of overtakes the entirety of meaning of your existence. Do you think, I know what it is for me. Do you have something like that, the broader vision that unites your work with Cardano and everything you've done in life, the big fish that you're going to end up in the dark of night struggling with in the last chapter of your life, [/INST] like the big problem you're taking on? Well, in mathematics, it was the Goldbach conjecture. I'd like to prove it. </s><s>[INST] That's probably not gonna happen. That's a really big fish. So do you still have a love for mathematics? [/INST] Oh God, yeah, of course I do. You never lose that, you never lose it. You lose the ability to do deep work and you don't have the creativity and the raw inspiration. This is why I've gotten to automated theory improving because what I lack, because I'm getting older, I can now have computers understand it and use AI to just solve this stuff. It's like the cheat codes for math, you know? So fuck those guys, we'll still do that. But you know, mathematics still has those last passions and there's all kinds of cool things that I'd like to see done. The quota complexes allow you to unify topology and number theory in really novel ways and there are all kinds of cool things, but who cares? Everybody has those white whales. Another thing I'd love to do is bring back the wooly mammoth. You know, that's George Church's hidden pleasure. In five to 10 years, it's actually gonna happen. And I have this beautiful ranch in Wyoming and I got to find a way to convince George to let me raise his clone mammoths fence on my ranch. Among the bison. Among the bison. They'd probably get along really well. And then I have to learn all these cool things about wooly mammoths. Like, you know, do you shave them in the summertime and let them get shaggy during the winter or do you just let that coat go and then fall? Who knows? </s> |
<s>[INST] it's very parallel, [/INST] it's very parallel among other things, and it's not obvious. I mean, this is a really kind of mind bending thing, because we've got to try and imagine where uh you know, we've got to try and imagine a parsing of the universe different from the one we have. And by the way, when we think about extraterrestrial intelligence and so on, I think that's kind of the key thing is, you know, we've always assumed, I've always assumed, okay, the extraterrestrials, at least they have the same physics. We all live in the same universe. They've got the same physics. But actually that's not really right, because the extraterrestrials could have a completely different way of parsing that the universe. So it's as if, you know, there could be for all we know right here in this room, you know, in the in the details of the motion of these gas molecules, there could be an amazing intelligence that we were like, but we have no way of we're not parsing the universe in the same way. If only we could pass the universe in the right way. You know, immediately this amazing thing that's going on in this huge culture that's developed in all that kind of thing would be obvious to us. But it's not because we have our particular way of passing the universe. </s><s>[INST] Would that thing also have agency? I don't know the right word to use, but something like consciousness, but a different kind of consciousness. [/INST] I think it's a question of just what you mean by the word, because I think that the, you know, this notion of consciousness and the okay, so some people think of consciousness as sort of a key aspect of it, is that we feel that the sort of a feeling of that we exist in some way, that we have this intrinsic feeling about ourselves? You know, I I suspect that any of these things would also have an intrinsic feeling about themselves. I've been sort of trying to think recently about constructing an experiment about what if you were just a piece of a cellular automaton, let's say, you know, what would your feeling about yourself actually be, and you know, can we put ourselves in the in the shoes, in the cells of the cellar or thomason, so to speak? Can we can we get ourselves close enough to that? That we could have a sense of what the world would be like if you were operating in that way? And it's a little difficult because, you know, you have to not only think about what are you perceiving, but also what's actually going on in your brain and our brains do what they actually do, and they don't, it's uh you know, I think there might be some experiments that are possible with, with uh you know, neural nets and so on, where you can have something where you can at least see in detail what's happening inside the system. And I've been sort of one of the, one of my projects to think about is is there a way of kind of uh kind of getting a sense kind of from inside the system about what its view of the world is and how it how it, you know, can we make a bridge to see? The main issue? Is this where, you know, it's a it's a sort of philosophically difficult thing because it's like we do what we do, we understand ourselves, um at least to some extent we humans understand ourselves. That's correct. And but yet okay, so what are we trying to do, for example, when we are trying to make a model of physics, what are we actually trying to do? Because, you know, you say, well, can we work out what the universe does? Well, of course we can we just watch the universe, the universe does what it does. But what we're trying to do when we make a model of physics is we're trying to get to the point where we can tell a story to ourselves that we understand that is also a representation of what the universe does. So it's this kind of, you know, can we make a bridge between what we humans can understand in our minds and what the universe does and in a sense, you know, a large part of my kind of life uh efforts have been devoted to making computational language, which kind of is a bridge between what is possible in the computational universe and what we humans can conceptualize and think about in a sense, what, you know, when I built well from language and our whole sort of computational language story, it's all about, how do you take sort of raw computation in this ocean of computational possibility and how do we sort of represent pieces of it in a way that we humans can understand and that map onto things that we care about doing. And in a sense, when you add physics, you're adding this other piece where we can, you know, mediated by computer, can we get physics to the point where we humans can understand something about what's happening in it? And when we talk about an alien intelligence, it's kind of the same story. It's like, is there a way of mapping what's happening? They're onto something that we humans can understand and, you know, physics in some sense, is like our exhibit one of the story of alien intelligence is it's a it's a, you know, it's an alien intelligence in some sense. And what we're doing in making a model of physics is mapping that onto something that we understand. And I think, you know, a lot of these other things that have, I've recently been kind of studying whether its molecular biology, other kinds of things, um which we can talk about a bit um the those are other cases where we're, in a sense, trying to again make that bridge between what we humans understand and sort of the natural language of that sort of alien intelligence in some sense, </s> |
<s>[INST] So figuring out how these concepts connect um you have a good idea of how to do that, [/INST] that's exactly what one of the problems that we are working on and and and and the way we are approaching that is basically saying okay you need to so the takeaway is that language is simulation control and your perceptual plus motor system is building a simulation of the world. Uh so so that's basically the way we are approaching it. And the first thing that we built was a controllable perceptual system and we built a schema networks which was a controllable dynamic system. Then we built a concept learning system that puts all these things together into programs as abstractions that you can run and simulate. Uh and now we are taking the step of connecting into language and uh it will be very simple examples. Initially it will not build the GPT three like examples but it will be grounded simulation based language </s><s>[INST] and for like the quarrying would be like question answering kind of thing. [/INST] And it will be in some simple world initially on, you know, but it will be about okay, can the system connect the language and grounded in the right way and run the right simulations </s><s>[INST] to come up with the answer and the goal is to try to do things that for example, G PG three couldn't do correct. Speaking of which if we could uh Talk about gpt three a little bit. I think it's an interesting thought provoking set of ideas that open your eyes pushing forward. I think it's good for us to talk about the limits and the possibilities and neural network. So in general, what are your thoughts about this recently released Very large 175 billion parameter language model? [/INST] So I have, I haven't directly evaluated it yet from what I've seen on twitter and other people evaluating it. It looks very intriguing, you know, I am, I am very intrigued by some of the property cities displaying and uh and of course the text generation part of that was already evident in GPT two, you know that it can generate coherent text or uh long distances that was uh but of course the weaknesses are also pretty visible in saying that okay, it is not really carrying a world state around um, and you know, sometimes you get sentences like I went up the hill to reach the valley or some, you know, completely incompatible statements or when you're traveling from one place to the other, it doesn't take into account the time of travel, things like that. So, those things I think will happen less than GPT three because it is trained on even more data and so and it does, it can do even more longer distance uh coherence, um but it will still have the fundamental limitations that it doesn't have a world model. Uh and it can't run simulations in its head to find whether something is true in the world or not. </s><s>[INST] Do you think within So it's taking a huge amount of text from the internet and forming a compressed representation, do you think in that could emerge something that's an approximation of a world model which essentially could be used for reasoning, it's I'm not talking about GPT three, I'm talking about GPT 45 and GPT 10. [/INST] Yeah, I mean, they will look more impressive than GPT three. So you can if you take that to the extreme them, uh a Markov chain of just fast order. And if you if you go to um I'm taking the other extreme, if you read Shannon's book, right? Uh he has a model of english text which is based on faster Markov chains. Second remark of change, the markets change and saying that, okay, the remark of chains look better than uh faster remark of change. So does that mean a faster Markov chain has a model of the world? Yes, it does. So, yes, in that level, when you go higher order models or more uh sophisticated structure in the model, like the transformer networks. Yes, they have a model of the text will um but that is not a model of uh the world. It's it's a model of the text world and it will have an interesting uh properties and it will be useful but just scaling it up, it's not going to give us a G. Or natural language understanding or meaning. </s><s>[INST] The question is whether being forced to compress a very large amount of text forces you to construct things that are very much like um because the ideas of concepts and meaning is a is a is a spectrum. [/INST] Uh </s><s>[INST] So in order to form that kind of compression, maybe it will be forced to figure out abstractions which look awfully a lot like the kind of things that we think about is uh as concepts as world models is common sense. Is that possible? [/INST] No, I don't think it is possible because the information is not there. </s><s>[INST] The information is uh is there behind the [/INST] text right now, unless somebody has written on all the details about how everything works in the world to the the absurd amounts like okay, it is easier to walk forward and backward uh that you have opened the door to go out of the things doctors were underwear, you know, unless all these things somebody has written down somewhere or you know, somehow the program found it to be useful for compression from some other text. Uh the information is not there. </s><s>[INST] So that's an argument that like text is a lot lower fidelity than the, you know, the experience of our physical world. Like [/INST] So it's worth 1000 words. Well </s> |
<s>[INST] reduced [/INST] Frankenstein Mac OS. And I mean that in the mess the best way it means it wasn't Mac os just changed a little. It was totally things were hacked out and changed and I think new code was inserted and it really was a whole set of things from all different places to make that first iphone Os. And then there was another team working on the apps and then another team working on the design of how it looked overall between all that stuff. So all of those things came together to create what we know as the first generation iphone </s><s>[INST] and those are all probably fascinating engineering challenges and great teams like the creating Frankenstein os, that's fascinating because you're simplifying simplifying but then you're just pulling different stuff from and you're basically inventing, I mean they're probably not thinking of it that way, but a new era of computing, a new kind of computer really is Frankenstein. [/INST] Right. And you didn't have to run mac software. If you look at some of the other smartphones of the time, like Windows and stuff, they were like, we need to make sure it runs Excel and it runs word or something like that and some reduced thing. This was like, no, no, no, no, this was born out of entertainment. So we didn't have to go and take all the same application, you know, all those other ones was about compatibility. This was about a whole new way of being, </s><s>[INST] What did you think about um, the steve jobs, presentation of the, the iphone, the sort of the first iphone, you know, [/INST] phone, internet communicator and the ipod. And </s><s>[INST] Then you're going to sort of present announcing three new products kind of thing. And then saying that it's all in one just this is a good example, one of the sort of historic presentations of a product. Clearly there's like some showmanship that works, some reason it works, it doesn't always work. It often doesn't work, but it did in this case it often did for steve like how did that feel? [/INST] What part </s><s>[INST] of the actually uh the design process was that presentation, You know what I mean? Like from the early because you said, so consider the why the press release at the very beginning, [/INST] steve was doing that the entire time he was working on that story from day one, he was pitching us this, this, this and then this and and then he would look at our faces because you wouldn't, most people wouldn't at least if you're working for him, I wouldn't tell him what you really thought of what he was saying, but he would look at your faces and then he would talk to a few real trusted confidence outside of the outside of the organization and see what they thought, right, and they could give him feedback on it and they could really challenge him, but he would also look at their faces and go and so when you see that, then he would modulate it and change it slightly and change it. So he was working during all of that time on the story and the storytelling, right? And the wise while we're working on that and helping us refine it just like the switch from plastic to glass, right all the time working on that. So when he comes out on stage, he does something that every marketer is told not to do Say these three things are now combined in one that is like the that they say that that is the laziest form of storytelling possible for </s><s>[INST] marketing, [/INST] right. Right. But it was the best one because it was all those pains. It was like I want my ipod but I want my communications and I want my internet browsing because I want it on the go. So I can look up things because it was information. And when you were on the road you had a laptop, you had an ipod and you had a phone that and you had to carry all of these things with you at once. Now we're gonna solve that pain for you and put it all together. So he was just showing you the pain and getting that virus of doubt and going, it's now in this one magical thing and he could come up and masterfully tell that story because he told it almost every day to all of these people inside very quietly. And then it was just Right, it was like a you know, a Tony Award winning play that had been working on for 10 years. But </s><s>[INST] it [/INST] was all that and of course he was dramatic at certain points and he would raise his voice and a wry smile or whatever </s><s>[INST] it was that [/INST] all those touches. He was an actor as well as a storyteller. </s><s>[INST] Yeah, [/INST] but but it was the truth, right? The truth came through. It was a nonfiction story and then he added those personal flourishes on top of it to for dramatic effect. It's amazing. </s><s>[INST] So there's a designer, you mentioned johnny av uh you both are brilliant designers, great human beings. There were some battles fought in the distant past between the two of you. Looking back, what is the positive characteristics of johnny that made you a better person? And designer having worked with him [/INST] watching the process that the design team that johnny lead, I don't know where, because that was over years, I didn't see all of those things. But watching the design process of really because it was really a team that was about materials, it was about form, it was about colors, it was about these physical characteristics. It it when we talked about this earlier was designed, what is design designs everywhere. Okay, so what they were really focused on was form, how the field was, how it looked, the aesthetics, the physical aesthetics and watching, going through that process. I learned so much in that process about how to do colors, how to do materials, how to think deeply about curves, right, and shadows and how it would look, not just in your hand, but how it would look in the photograph you're gonna take from marketing, </s> |
<s>[INST] The ad reads today are great. So simply safe is the protection of physical space. Expressly pianist. The initial protection in digital space. How cool would it be if there is not like a hybrid physical digital space. And then we have tools that we can carry from the physical to the digital and back. The protection will come along with us. And of course, if the anarchists have their say that will be provided by private companies and we'll compete over their customers and through that process of capitalism would then create the best product and the most affordable product that is if the anarchists have their way Michael malice has entered the chat anyway, go to simply safe dot com slash legs to customize your system and get a free security camera. Yes, friends. I said free again, that's simply safe dot com slash lex. And now here's my conversation with cal Newport's what is deep work. Let's start with a big question. [/INST] So, I mean, it's it's my term for when you're focusing without distraction on a cognitive demanding task, which is something we've all done. But we had never really given it a name necessarily that was separate from other type of work. And so I gave it a name and said, let's compare that to other types of efforts you might do while you're working and see that the deep work efforts actually have a huge benefit that we might be underestimating </s><s>[INST] what does it mean to to work deeply on something. [/INST] You know, I have been calling it hard focus in my riding before that. Well, so the context you would understand I was in the theory group in C sale at MIT. Right? So I was surrounded at the time when I was coming up with these ideas by these professional theoreticians and that's like a murderer's row of thinkers. They're right. I mean it's like turing award turing award, macarthur turing award. I mean, you know the crew, right? Theoretical computer science. Theoretical computer science. Yeah. Yeah. So so I'm in the theory group right? Doing theoretical computer science. And I published a book. So you know, so I was in this milieu whereas being exposed to people, We're focused was their tier one skill. Like that's what you would talk about right? Like how how intensely I can focus. That was the Wiki Skill. It's like you're 4:40 time or something if you were an athlete. Right? </s><s>[INST] So this is something that people were actually the theory folks are thinking [/INST] about really </s><s>[INST] like they're openly discussing. Like how do you focus? [/INST] I mean, I don't know if they would, you know, quantify it but but focus was the tier one Skill. So you you would come in here would be a typical that you'd come in uh and eric domain would be sitting in front of a white board right? With a whole group of visitors who had come to work with them and maybe they projected like a grid on there because they're working on some graph theory problem. You go to lunch, you go to the gym, you come back. They're sitting there staring at the same same white board, right? Like that's the tier one skill. This is the </s><s>[INST] difference between different disciplines. Like I I often feel for many reasons like a fraud but I definitely feel like a fraud. When I hang out with like either mathematicians or physicists, it's like, it feels like they're doing the legit work because when you talk closer and computer science to get to programming or like machine learning, like the the experimental machine learning or like just the engineering version of it, it feels like you're gone so far away from what's required to solve something fundamental about this universe. It feels like you're just like cheating your way into like some kind of trick to figure out how to solve a problem in this one particular case. That's how it feels like it's uh I'd be interested to to hear what you think about that because um programming doesn't always feel like you need to think deeply to work deeply. But sometimes it does, it's a weird dance [/INST] for sure code does right? I mean, especially if you're coming up with original algorithmic designs. I think it's a great example of deep work. I mean yeah, the hardcore theoreticians, they push it to an extreme. I mean I think it's like knowing that athletic endeavor is good and then hanging out with an olympic athlete like, Oh I see that's what it is now for the grad students like me, we're not anywhere near that level, but the faculty, the faculty in that group, these were the cognitive olympic athletes, but Cody and I think is a classic example of deep work because I got this problem I want to solve, I have all of these tools and I have to combine them somehow creatively and on the fly, but so basically I had been exposed to that. So I was used to this notion when I was in grad school and I was writing my blog, I write about hard focus, you know, there was a term I used, Then I published this book, so good, they can't ignore you, which came out in 2012. So like, right, as I began as a professor and that book had this notion of skill being really important for career satisfaction, that it's not just following your passion, you have to actually really get good at something and then you use that skills as leverage. And there's this big follow up question to that book, Okay, well how do I get really good at that? And then I look back to my grad school experience, I was like, huh, there's this focus thing that we used to do, I wonder how generally applicable that is into the knowledge sector. And so as I started thinking about it, it became clear there is this interesting storyline that emerged, okay, actually undistracted concentration is not just important for esoteric theoreticians, it's important here and support here and support here and that involved to the deep work hypothesis, which is across the whole knowledge work sector, focus is very important and we've accidentally created circumstances where we just don't do a lot of it. </s> |
<s>[INST] invisibility, right? There's some aspect why Fauci was very interesting. In my recent memory, there's not been many faces of scientific policy, of science policy. He became the face of that. And, you know, as there's some of it is meme, which is, you know, basically saying that he is science or to some people represent science, but in the... Or quote unquote science or whatever, yeah. Yeah. The positive aspect of that is that there is accountability when there's a face like [/INST] that. He was a public face. Yeah, right. But you can also see the Fauci example shows you why a lot of these folks do not want to be public because they enter a political, you know, immediate minefield. I'm actually sympathetic to that aspect of it. What I'm not sympathetic to is the concept that in 2022 that the unelected, unfireable, anonymous American regulator should be able to impose regulatory policy for the entire world. We are not the world of 1945, you know, it is not something where these other countries are even consciously consenting to that world. Just as give an example, you know, there's a concept called challenge trials. Okay. The Moderna vaccine was available very, very early in the pandemic, you can just synthesize it from the sequence. And challenge trials would have meant that people who are healthy volunteers, okay, they could have been soldiers, for example, of varying ages who are there to take a risk their lives for their country, potentially, okay, it could have been just healthy volunteers, not necessarily soldiers, just patriots of whatever kind in any country, not just the US, but those healthy volunteers could have gone. And at the early stages of the pandemic, we didn't know exactly how lethal it was going to be because, you know, Li Wenliang and, you know, 30 year old, 30 somethings in China were dying from this, it seemed like it could be far worse. How lethal the virus would be. Yeah, it may be, by the way, that those who are the most susceptible to the virus died faster earlier. It's as if you can imagine a model where those who were exposed and had the lowest susceptibility also had the highest severity and died in greater numbers early on. If you look at the graph, like deaths from COVID were exponential going into about April 2020, and then leveled off to about 7,500 10,000 a day, and then kind of fell, right? But it could have gone to 75,000 at the beginning. So we didn't know how serious it was. So this would have been a real risk that these people would have been taking. But here's what they would have gotten for that. Basically, in a challenge trial, somebody would have been given the vaccine, and then exposed to the virus and then put under observation. And then that would have given you all the data, because ultimately the synthesis of the thing, I mean, yes, you need to scale up synthesis and manufacturing and what have you. But the information of whether it worked or not, and it was safe and effective. And like that could have been gathered expeditiously with </s><s>[INST] volunteers for challenge trials. And you think there'd be a large number of volunteers? [/INST] Absolutely. What's the concern there? Is there an ethical concern of taking on volunteers? Well, so let me put it like this. Had we done that, we could have had vaccines early enough to save the lives of like a million Americans, especially seniors and so on. Okay, soldiers, and we're generally first responders and others, you know, I do believe there's folks who would have stepped up, you know, to take that risk. Heroes walk among us. Yeah, yeah, that's right. Like if military service is something which is ritualized thing, people are paid for it, but they're not paid that much, they're really paid in honor, you know, and in duty and pictures, that is actually the kind of thing where I do believe some some fraction of those folks would have raised their hand for this important, you know, task. I don't know how many of them, but I do think that we were the volunteers would have been there. There's probably some empirical test of that, which is a, there's a challenge trials website, there's a Harvard prof who put out this proposal early in the pandemic, and he could tell you how many volunteers he got. But something like that could have just shortened the the time from pandemic to functional vaccine, right, to today's, even if you'd actually really acted on it, the fact that that didn't happen, and that the Chinese solution of lockdown, that actually, you know, at the beginning, people thought the state could, you know, potentially stop the virus, stop people in place, it turned out to be more contagious than that. Basically, no NPI, no non pharmaceutical intervention, really turned out to work that much, right. And actually, at the very beginning of the pandemic, I said something like, Look, is actually February 3, about a month before people, you know, I was just watching what was going on in China, I saw that they were doing digital quarantine, like using WeChat codes to like block people off and so on. I didn't know what was going to happen, but I said, Look, if the coronavirus goes pandemic, and it seems it may, the extreme edge case becomes the new normal. It's every debate we've had on surveillance, deplatforming and centralization accelerated pandemic means emergency powers for the state, even more than terrorism or crime. And sometimes a solution creates the next problem. My rough forecast of the future, the coronavirus results in quarantines, nationalism, centralization. And this may actually work to stop the spread. But once under control, states will not see their powers. So we decentralized. And I didn't know whether it was going to stop the spread. But I knew that they were going to try to do it. Right. And, you know, look, it's it's hard to call every single thing right. And I'm sure, you know, someone will find some errors. But in general, I think that was, that was actually pretty good for like early February of 2020. Right. So it's my point, though. The point is, rather than copying Chinese lockdown, what we should have had were different regimes around the world, to some extent, Sweden defected, you know, from this, right? They had like no lockdowns or what have you. But really, the axis that people were talking about was lockdown versus no lockdown. The real axis should have been challenge trials versus no challenge trials. We could have had that in days. Okay. And that those are two examples on both vaccines and </s> |
<s>[INST] talk to to Vladimir [/INST] Putin once again. Now, what kind of, what kind of things </s><s>[INST] would you talk about [/INST] here. What kind of questions </s><s>[INST] would you ask? Huh? [/INST] Well One thing I would certainly ask is what you were thinking on February 23 and I would ask him to reply to my question about what if you took this to phase two, You you surrendered in Donbass, You know, no ego about it. You </s><s>[INST] just surrendered. It's in your [/INST] interest to your country. And you invited all the refugees from Donbass into Russia as much as they can. What would you do </s><s>[INST] now? What's [/INST] the what's the U. S. Next move And in your opinion? How are you going to? Okay, where are we gonna go? That's that would be the key question because it's but he didn't go that way. He chose to take </s><s>[INST] the sanctions [/INST] and and to go this way. Why he did? That is a key question for our time. Uh, perhaps it was a mistake. Perhaps it was his judgment perhaps as I said. But I don't knowing the man I did. I </s><s>[INST] don't think so. I [/INST] think it was calculated. Now. </s><s>[INST] This is projection [/INST] and speculation. </s><s>[INST] But there's something different [/INST] about him in </s><s>[INST] the past [/INST] several months. </s><s>[INST] It could be the covid thing, the [/INST] isolation that you mentioned. </s><s>[INST] I listened to a lot of interviews and speeches in [/INST] Russian. And </s><s>[INST] there's uh, [/INST] there's </s><s>[INST] something about power [/INST] over time that </s><s>[INST] can change you. That can isolate you. Well, [/INST] When I was there now, he'd been in office for already 15 years. He had power. He didn't miss use it in my opinion. He was very, even I saw him go on television and talked to his fellows the same way. He always talked to them. He grew with it. He grew in intelligence and knowledge because he had he had dealings with the whole world. Now people had come to him. He was very well known in Africa and Middle East. Certainly Syria and I I just never saw misuse of his power. I saw humility in him actually. </s><s>[INST] So perhaps [/INST] there was a calculation and </s><s>[INST] he calculated wrong [/INST] in terms of what </s><s>[INST] happens if he doesn't invade? [/INST] Perhaps there was a </s><s>[INST] calculation. Perhaps he had [/INST] a calm and clear mind and he calculated wrong. Well, he also made the point that he had the talk of Zelensky saying nuclear weapons, we're going to come into Ukraine. There was talk about that right before the invasion to certainly that would have set off alarms. You know, the United States is already </s><s>[INST] kind of doing that [/INST] by not only putting its intelligence and its heavy weaponry into Ukraine, but </s><s>[INST] you've got to deal with the [/INST] question. The next question that comes up, the most immediate question </s><s>[INST] is, [/INST] is the United States, </s><s>[INST] I'm going to [/INST] start and I'm saying this is great. They are making a lot of noise. The United States. The press about Russia using nuclear weapons and chemical weapons. That's a lot of noise. Again, going back to my </s><s>[INST] analogy, when the [/INST] United States starts that it starts the conversation going, it's in the interests of the United States for Russia to be pinned with any kind of chemical or </s><s>[INST] nuclear [/INST] uh incident. Exactly, for example, would be very not simple, but it would be possible to explode a nuclear device in Donbass in Donbass and kill thousands of people. And we would not know right away who did it. But of course the blame would go right to Russia, right to Russia. Even if it didn't make sense, If there was no motivation for </s><s>[INST] it, it would just be [/INST] blamed on Russia. The United States might well be the one who does that false flag </s><s>[INST] operation. It would not be [/INST] beyond them. They would it would be a very dramatic solution to sealing this war off as a major victory for the United States. That's </s><s>[INST] terrifying. But it can happen. [/INST] It can happen in one kiloton device, low yield this. It's possible. So when you </s><s>[INST] walk across that line, you can potentially never walk back. [/INST] Well, I think the United States is calculating that it's a dangerous. Yes, I agree. But I think the neoconservative arrogance is such that they really believe they can push their advantage to the max </s><s>[INST] now [/INST] because of all these propaganda </s><s>[INST] successes up to now, [/INST] the Ukrainian army could be wiped out. For all we know there's all that's left is or neo nazi brigades, but they're being advised very well by us and they're sending the weapons in a huge amounts of weapons. What about american budget? No one talks about how much money we're giving to Ukraine. We were it's a billion dollars already in weaponry and not most of it just poured in. What about, You know, the the Russian budget is uh, defense budget is 60 some billion dollars a year. It's nothing compared to the </s><s>[INST] United States 1 15th of [/INST] it. But yet we we've put so much weaponry into Ukraine. The money we've spent on Ukraine is all equivalent almost to what we've spent on Covid in in our own country. It's astounding the distortion of our priorities. </s><s>[INST] Uh, [/INST] there's also chemical, don't forget chemical is probably the easier way to go. But in Syria there was far too many incidents of, of America in its quest to </s><s>[INST] demonize [/INST] assad. And the Russians of all these chemical attacks that were happening that they were vowing came from </s><s>[INST] Russia. And [/INST] uh, in spite of the </s><s>[INST] fact that Russia has [/INST] pulled out of the, </s><s>[INST] signed the agreement [/INST] on Chemical arms and not and apparently destroyed its stock several years ago. </s><s>[INST] It's it's [/INST] strange that the strangest incidents </s><s>[INST] happened in Syria, [/INST] you go back to trace </s><s>[INST] everyone. [/INST] Good journalism was done. The white helmets got a lot of fame but they were corrupted and many good journalists tried to point out the inconsistencies in the american accusations, robert Perry among them who was one of my mentors at consortium press a lot of good journey. You'd have to go back but trace each like you would trace each time you </s> |
<s>[INST] I I tend to believe it's everywhere except for the fact that we don't even know how to define what life is. Like what is everywhere exactly what we're talking about. It's it's very possible that there's not anywhere in the universe. An organism with two legs and two arms With two eyes and mostly hairless walking around at this time scale. But like there could be a very different kind of other things. Um It was interesting that some people this is not a common [/INST] belief but a </s><s>[INST] friend named lee Cronin, he's a chemist and biologist. And he believes that if we ran evolution over and over and over and over on earth it would get very different. Not just you wouldn't just get different organisms. You get very different biology. That's a weird thing. I mean most people kind of assume, well it kind of you know it fits to the environment and you're going to get similar things. Maybe not humans and so on. But to get very different biology. Like starting from the bacteria to the you know just how well [/INST] the idea that it would be DNA based </s><s>[INST] DNA based on some [/INST] other planet. That seems to me like saying they're speaking Swahili on some other planet. I mean the odds of that particular architecture I think are infinitesimally small. </s><s>[INST] What's the coolest stone you've ever seen? [/INST] Oh my God there's so many. And what </s><s>[INST] what what defines is it rarity? Is it just raw beauty? What what captivates your [/INST] eye like a storied stone? I have a very beautiful Fairburn maggot which is multiple layers. And there's something I call agape paralysis because to polish it you have to go through the layers which means you're destroying the layers. And maybe what should be done is it should be like a movie where you you film the entire process of cutting and polishing so that it's not dead. In other words, what was the diamond when it started rough the rough diamond is gone. But if you could sort of to a filmic version of a cutting process so that the the stone would exist in a from a pre polished to a polish state all as a kind of enough tea or so. Another. So the other thing I fantasized about is how pattern recognition technology will probably in the future allow us to discover all kinds of amazing stones including for example fossil skulls, fossil skulls of humans. Now it's kind of a chance process that you discover a skull in East Africa. But why not have a drone moving constantly scanning for pattern recognition of human skull. Human teeth very slowly and then </s><s>[INST] on the surface, just [/INST] Above the surface, just 10 ft above the surface. 20 ft above the surface. </s><s>[INST] I'm sorry. You think you'll be able to find skulls on the surface? [/INST] Yes. Yes. In the middle of a place that no one has looked in. These areas are vast. Right? So it could be found on the service. Then move to the next layer, then find it under the surface as well. There's lidar, there's all kinds of ways. We're finding jungles and jungle cities in under the amazon that people didn't know about. </s><s>[INST] Do you think there's something out there that would just blow your mind? [/INST] Oh, for sure. For </s><s>[INST] sure. [/INST] Yeah, no doubt. </s><s>[INST] Oh, man. And uh, how much of it is a little bit underground? Right. Or how much of it is in the ocean? [/INST] I mean here, right, here we are in the bay area. We know that much of the native american civilization here was under the bay Because 6000 years ago the Bay was dry. It was a river not obey. And so all of those, whatever material culture archaeological traces existed, there are now at least preserved under the water. So I think we're just beginning to touch the could </s><s>[INST] be treasure too. I mean like literally like you said, we lose the wisdom or we'll lose the knowledge. But I mean, well if the if there's the pyramids, right, It's the great wonders of the world, there might be other wonders that are completely lost. Just I [/INST] mean, one of the stones you asked about stones, I like I I like stones for example, every now and the dinosaurs would eat rocks as gizzard stones and then you find them in their guts, in their bones. Well, every now and then they would eat a piece of petrified wood. So the idea that something was a tree and then stone and then swallowed by a dinosaur and ground up in the gizzard and polished and then left in a, you </s><s>[INST] know, so I [/INST] like things that have been through dramatic. </s><s>[INST] There's a story [/INST] there. </s><s>[INST] Yeah, I mean that that's okay. The really fascinating thing why seeing Allah or crosses in the stone is it feels like the stone has wisdom because it has been there's a there's a it's been through so many generations of humans. It's like bigger, [/INST] Right? </s><s>[INST] It's seen it all [/INST] also it's also the intellectual question of intelligent design. In other words, When people say intelligent design, mostly it's bogus. But there are several interesting examples of actual intelligent design meaning when is a stone? The product of artifice? And when is it a geo fact produced by nature? And that's that was an important discovery in the 19th century. The zone of percussion, It's called the the percussion zone or how do you know that a signal from out of space is an intelligent signal and as opposed to hydrogen doing something or some natural thing. That's the genuine problem of intelligent design? How do you know if it's pie maybe if it's e if it's some, you know, pattern, how do you know that that's an intelligent signal. How do you know that an artifact in the ground is, you know, we'll see in the clouds of a face, it's called paradise oily A we have a kind of a built in ability to see faces where they really aren't there? Right. That's why kids like clowns, you know, we've evolved, that's a baby's evolved to recognize their parents and so forth. But when is it a projection? And when is it really in the in the in the stone? And that was a big question with the rise of fossils. You know, if you find a curly thing, is that life or is it non life? Is it? You know, people have made this mistake before, you know, they'll find a rock on on the moon or mars. They say, oh, this is a face or whatever. Well, you know, no, that's that's just projection. That's paradise Alia, </s> |
<s>[INST] So, yes, and I just spoken with Gnome chomsky who basically was part of uh dismantling the behaviors movement, but and I'm with that 100% for studying human behavior, but I am one of the few people in the world who has made room bas scream in pain and uh I just don't know what to do with that empirical evidence because it's hard to sort of philosophically, I agree, but uh the only reason I philosophically agree in that case is because I was the programmer. Well, if somebody else was a programmer, I'm not sure I would be able to interpret that well. So it's uh I think it's a new world that uh I was just curious what your thoughts are for now you feel that the display of the what we can kind of intellectual say is a fake display of suffering is not suffering. [/INST] That's right. That would be in my view, but that's consistent of course with the idea that it's part of our nature to respond to this display if it's reasonably authentically done uh and therefore it's understandable that people would feel this. Um and maybe as I said, it's even a good thing that they do feel it and you wouldn't want to harden yourself against it because then you might harden yourself against beings who are really suffering. </s><s>[INST] But there's this line, you know, So you said once a artificial general intelligence system, a human level intelligence system become conscious, I guess if I could just linger on it now, I've wrote really dumb programs to just say things that I told them to say, but how do you know when uh when a system like Alexa was just officiate complex that you can't introspect the hot works, starts giving you signs of consciousness through natural language. That there's a there's a feeling there's another entity there that's self aware that has a fear of death, immortality, that has uh awareness of itself that we kind of associate with other living creatures. It I guess I'm sort of trying to do the slippery slope from the very naive thing where I started into into something where it's sufficiently a black box to where it's starting to feel like it's conscious, Where is that threshold or you would start getting uncomfortable with the idea of robots suffering? Do you think? [/INST] I don't know enough about the programming that we go into this really to answer this question, but I presume that somebody who does know more about this could look at the program and see whether we can explain the behaviors in a parsimonious way that doesn't require us to suggest that some sort of consciousness has emerged, or alternatively, whether you're in a situation where you say, I don't know how this is happening, the program does generate a kind of artificial general intelligence which is autonomous, starts to do things itself and is autonomous of the basic programming that set it up. And so it's quite possible that actually we have achieved consciousness in a system of artificial intelligence </s><s>[INST] sort of the the approach that I work with most of the community is really excited about now is with learning methods. So machine learning and the learning methods are unfortunately are not capable of revealing. Which is why somebody like no child criticizes them. You have created powerful systems that are able to do certain things without understanding the theory, the physics, the science of how it works. And so it's possible if those are the kinds of methods that succeed, we won't be able to no exactly sort of try to reduce uh try to find whether there this thing is conscious or not. This thing is intelligent or not. It's simply giving when we talk to it it displays wit and humor and cleverness and emotion and fear. And then we won't be able to say where in the billions of nodes neurons in this artificial neural network is the fear coming from sort of in that case. That's a really interesting place where we do now start to return to behaviorism and say [/INST] yeah, that's that's that is an interesting issue. Uh I would say that if we have serious doubts, I think it might be conscious, then we ought to try to give it the benefit of the doubt. Just as I would say with animals, we I think we can be highly confident that vertebrates are conscious but when we get down and some invertebrates like the octopus. But but with insects it's much harder to be to be confident of that. I think we should give them the benefit of the doubt where we can, which means, you know, I think it would be wrong to torture an insect, but this doesn't necessarily mean it's wrong to sniper mosquito that's about to bite you and stop you getting to sleep. So I think you you try to achieve some balance in these circumstances of uncertainty </s><s>[INST] if it's okay with you, If you can go back just briefly. So 44 years ago, like you mentioned, 40 plus years ago, you've written animal liberation, the classic book that started that launched it was the foundation of the movement of Animal liberation. Did can you summarize the key set of ideas that underpin that book? [/INST] Certainly the key idea that underlies that book is the concept of species is um which I did not invent that term. I took it from a man called Richard rider who was in Oxford when I was and I saw a pamphlet that he'd written about experiments on chimpanzees that used that term. But I think I contributed to making it philosophically more precise and to getting it into a broader audience. And the idea is that we have a bias or prejudice against taking seriously the interests of beings who are not members of our species just as in the past, Europeans, for example, had a bias against taking seriously the interests of Africans racism and men have had a bias against taking seriously the interests of women. Sexism. So, I think something analogous, not completely identical, but something analogous goes on, and has gone on for a very long time with the way humans see themselves visibly. Animals. We see ourselves as more important. We see animals as existing to serve our needs in various ways. And you can find this very explicit in earlier philosophers from Aristotle through the account and others. Uh and um either we don't need to take their interests into account at all, or we can discount it because they're not humans. They can a little bit, but they don't count nearly as much as humans do. Uh my book, I used that that attitude is responsible for a lot of the things that we do to animals that are wrong, confining them indoors, and very crowded, cramped conditions in factory farms to produce meat or eggs or milk more cheaply using them in some research that's by no means essential for survival or well being a whole lot. You know, some of the sports and things that we do to animals. So, I think that's unjustified, because I think uh the significance of pain and suffering does not depend on the species of the being who is in pain or suffering any more than it depends on the race or sex with the being who is in pain or suffering. And I think we ought to rethink our treatment of animals along the lines of saying if the pain is just as great an animal then it's just as bad that it happens as if it were a human. </s> |
<s>[INST] But I think you have to think about that right [/INST] now. You can I can very easily see him in the room being very calm and charming. And then you think, okay, the guy with the speeches is the act and he's putting on a show for his people and this is the real one. </s><s>[INST] Okay, so let's let's take somebody as an example. Let's take uh our mutual friend Vladimir Putin, Okay, I don't know why he's saying his name, makes my voice crack [/INST] because you're scared. You can hear you like Beetlejuice </s><s>[INST] Vologda. So there's a lot of people [/INST] that you're the one who built, You </s><s>[INST] know, that was that was a collaboration. Um what's it's a double blind engineering effort where I was not told of who my maker was, there's a back story, but [/INST] there's a talking cricket Pinocchio mm </s><s>[INST] the real voice and I talked about him quite a bit because I find him fascinating. Now there's there's a really important line that people say, like why does lex admire Putin? I do not admire Putin. I find the man fascinating. I find Hitler fascinating. I find a lot of figures in history fascinating, both good and bad and the figures just as you said that are with us today, like Vladimir Putin, like donald trump like Barack Obama. It's difficult to place them on the spectrum of good and evil because that's only really applies to like when you see the consequences of their action is a historical context. So there's some people who say that Vladimir Putin is evil and based on our discussion about Hitler, that's something I think about a lot, which is in the room with Putin And there's also a lot of historical descriptions of what it's like to be in the room with Hitler in the 1930s. There there is a lot of charisma in the same way I find Putin to be uh very charismatic in his own way, the humor, the wit, the brilliance. The there's um there's a simplicity of the way he thinks that really if taken at face value looks like a very intelligent, honest man thinking practically about how to build a better Russia constantly, almost like, um that's an executive, like he loves, he looks like a man who loves his job in a way that trump for example, doesn't meaning like he loves laws and rules and how to, [/INST] there's no adversarial press, so that's going to help and he's popular with his people. That's also gonna help enormously. </s><s>[INST] I'm talking about strictly the man directly the words coming out of his mouth, like all the videos interviews I watched, I'm based on that, not the press, not the reporting. You can just see that here's a man who is able to display a charisma that's not like I can see that's why I love joe Rogan is like, you could tell the guy is genuine and there's a good person and you can tell immediately that like, once you meet joe that he's going to be offline, also good person, you can tell, there's like signals that we send that are like difficult to kind of describe in the same way, you can tell Putin is like, he genuinely loves his job and wants to build a better Russia. There's the argument that he is action, evil man behind that charisma or is able to, you know, assassinated people, Uh, you know, limit, free press, all those kinds of things like that's what do we do with that? So what do human beings like journalists or what do other leaders when they're in the room with Putin do with those kinds of notions in deciding how to act in this world and deciding what policy to enact all those kinds of things. Just like with Hitler, when chairman is in the room with Hitler, how does he decide how to act? [/INST] Well, let's go back to like my wheelhouse, which is north Korea right, so uh when your entire world is based on being against trump and everything trump does is buffoonery or or kind of productive, the conclusion of your reporting is going to be pretty much given, I was very hopeful that there would be some positive outlooks or outcomes rather of trump's meeting with kim jong un. Um it looked like there was a space for things to go a bit better. I talked about a lot of the time um and trump was under no illusions about who he was dealing with. Um People pretend that oh he was kind of naive, he had one of the refugees that has state of the union, you know, lifting up his crutch. The first thing he sat down and talked to Jinping about in mar a Lago right after he became inaugurated was North Korea. Barack Obama said that when he sat down trump in the White house during the transfer of power, he said North Korea is the biggest issue. So I think a good leader whether or not you consider trump a good leader has to be aware of. All right, I'm going to have to have relationships of some kind, even if it's adversarial with some really evil evil horrible people, which kim jong un clearly is. </s><s>[INST] Well, I don't think there's anybody that has a perspective that north north korean, kim jong un or ill are not evil, right? But with In 1930s, Germany, isn't it a little bit more nuanced [/INST] because Hitler hasn't done anything yet and he's just to blow hard and he's an anti Semite shore. </s><s>[INST] But he's what about like before the war breaks out? Like what about the basic uh actionable anti Semitism when you're like just attacking hurting, which [/INST] was Kristallnacht or talk about the night of </s><s>[INST] long knives, kristallnacht. So it's the night of the broken glass [/INST] with long knives when he assassinated a bunch of his people. That was something different. </s> |
<s>[INST] That's a good line intelligible intelligence. [/INST] Yes. And that we shouldn't settle for something that seems intelligent, but we should it should be intelligible. So that we actually trust it because we understand it right? Like again, Ellen trusts his rockets because he understands Newton's laws and thrust and everything works. Uh, and let me tell you what can I tell you why? I'm optimistic about this? I think I think there is we've made a bit of a mistake where we, some people still think that somehow we're never going to understand neural networks and we're just gonna have to learn to live with us. It's this very powerful black box basically for those. You know, I haven't spent time building their own. It's super simple what happens inside you send in a long list of numbers and then you do a bunch of operations on them? Multiply by matrices, etcetera, etcetera. And some other numbers come out the output of it. And then there are a bunch of knobs you can tune and when you change them, you know, it affects the computation, input, output relation, and then you just give the computer some definition of good and it keeps optimizing these knobs until it performs this good as possible. And often you go like, wow, that's really good. This robot can dance or this machine is beating me at chess now. And in the end you have something which, even though you can look inside it, you have very little idea of how it works. You know, you can print out tables of all the millions of parameters in there. Is it crystal clear now how it's working? And of course not. Right. So many of my colleagues seem willing to settle for that and I'm like, no, that's like the halfway point. Uh some have even gone as far as sort of guessing that the mr the inscrutability of this is where the some of the power comes from and sort of some sort of mysticism. Uh I think that's total nonsense. I think the real power of neural networks comes not from inscrutability, but from differential ability and what I mean by that is simply that the output depends changes only smoothly if you tweak your knobs and then you can use all these powerful methods we have for optimization in science. We can just tweak them a little bit and see, did that get better or worse? That's the fundamental idea of machine learning that the machine itself can keep optimizing until it gets better. I suppose you wrote this and algorithm instead in python or some other programming language. And then what what the knobs did was they just changed random letters in your in your code. Now we'll just epically failed. Right? You change one thing and instead of saying print that says sint syntax error, you don't even know was that this for the better or for the worst. Right. This to me is this is what I believe is the fundamental power of neural networks. </s><s>[INST] And just to clarify the changing of the different letters in the program would not be a differential process. [/INST] It would make it in an invalid program typically. And then you wouldn't even know if if you change more letters, if it would make it work again. Right. </s><s>[INST] So that's the magic of neural networks. Uh, the [/INST] inscrutability, the defense ability that that every every setting of the parameters is a program. And you can tell it is it better or worse? Right. And so </s><s>[INST] so you don't like the poetry of the mystery of neural networks as the source of its [/INST] power. I generally like poetry, but not in this case. It's so misleading. And above all, it shortchanges us. It's failed. It makes us underestimate what we can the good things we can accomplish because so what we've been doing in my group is basically step one, train the mysterious neural network to do something well. And then step to do some additional Ai techniques to see if we can transform this black box into something equally intelligent that you can actually understand. So for example, I'll give you one example this ai fireman project that we just published. Right, So we I took the 100 most famous or complicated equations from one of my favorite physics textbooks. In fact the one that got me into physics in the first place, Feinman lectures on physics and so you have a formula. You know, maybe it has what goes into the formula is six different variables and what comes out is one. So then you can make like a giant Excel spreadsheet with seven columns, you put in just random numbers for the six columns for those six input variables. And then you calculate with the formula of the seventh column output. So maybe it's like the forest equals in the last column some function of the other. And now the task is okay if I don't tell you what the formula was, can you figure that out from looking at my spreadsheet? I gave you this problem is called symbolic regression. If I tell you that the formula is what we call a linear formulas. So it's just that the output is sum sum of all the things input the time, Some constants. That's the famous easy problem. We can solve we do it all the time in science and engineering. But the general one if it's more complicated functions with logarithms or co signs or other math. It's a very very hard one and probably impossible to do fast in general just because the number of formulas with and symbols just grows exponentially just like the number of password you can make grow dramatically with length. So so we but we had this idea that if you first have a neural network that can actually approximate the formula, you just train that even if you don't understand how it works that can be the first step towards actually understanding how it works. So so that's what we do first and then we study that neural network now and put in all sorts of other data that wasn't in the original training data. And use that to discover simplifying properties of the formula. And that lets us break it apart often into many simpler pieces in a kind of divide and conquer approach. They were so we were able to solve all of those 100 formulas, discover them automatically plus a whole bunch of other ones. And uh it's a it's actually kind of humbling to say that this code which anyone who wants now is listening to this. You can type pip, install a I find men on the computer and run it. You know, it can actually do what Johannes kepler spent four years doing when he stared at MARS data until he's like funny Eureka. This is an ellipse. Uh this will do it automatically for you in one hour, right? Or max plank. He was looking at how much radiation comes out from different wavelengths from a hot object and discovered the famous black body formula. This discovers it automatically. Uh huh. I'm I'm actually excited about yeah. Seeing if we can discover not just old formulas again, but new formulas that no one has seen before. </s> |
<s>[INST] put it in perspective. I'm not sure a lot of people understand the deep human side or even the military strategy side of this war. There's a lot of experts outside of the situation there commenting on it with certainty and that kind of hurts me because I feel like there's a lot of uncertainty. There's so much propaganda, it's very difficult to know what is true. Um Yeah, so, so my whole hope was to travel to Ukraine to travel to Russia, to talk to soldiers, to talk to leaders, to talk to real people that have lost homes, that have lost family members that who this war has divided, who this war changed completely how they see the world. Uh whether they have love or hate in their heart to understand their stories. Um, I've learned a lot on the human side of things by having talked to a lot of people there, but it has been on the Ukrainian side, for me, currently traveling to the Russian side is more difficult. Let me ask you about your now friend, can we go as far as to say his friend in a Sabu Hector Monsegur what's the what's what's the story? What's your long story with him? Can you tell me about? What is law? Sec who is Sabu and who's anonymous? What is anonymous? Where's the right place to start? That story? [/INST] Probably anonymous, anonymous is a was a decent still is, I guess a decentralized organization. Uh they call themselves headless. But once you look into them a little ways and they're not really headless. Um the power struggle comes with whoever has hacking ability. Um that might be your good hacker or you have a giant bottom that used for d does. Um So so you're gonna wield more power if you can control where it goes anonymous started doing their like hacktivism stuff in 2010 or so. Um The word hack was in the media all the time then. Um and then right around then there was a federal contractor named H. B. Gary Federal. The Ceo is Aaron Barr and Aaron Barr said he was gonna come out and be anonymous is anonymous. He's gonna come out and talk at black hat or def con or one of those and say you know who they are, He figured it out. So he figured it out by based on you know when people were online when people in IRC when tweets came out it was there was no scientific proof behind it or anything. So he's just gonna falsely name people that were that were in anonymous. So anonymous went on the attack. They went and hacked in hp. Hp A. Federal and they turned his life upside down. They took over his twitter account and all that stuff. Um pretty quickly. So </s><s>[INST] I have very mixed feelings about all of this. [/INST] Okay I </s><s>[INST] get like part of me admires the positive side of the activism. Is there no room for admiration there of the funk. You to the man. [/INST] Not at the time again there was a violation of 18 U. S. C. 10 30. So it was my job you know so at the time though in retrospect shore. But </s><s>[INST] what was the philosophy of the activism was what the philosophically were they at least expressing it for the good of humanity or no [/INST] they outwardly said that they were going to go after people that they thought were corrupt. So they were judge and jury on corruption. They were to go after it once you get inside and realized what they were doing they were going after people that they had an opportunity to go after. Um So maybe someone had a zero day and then they searched for servers running that zero day. And then from there let's find a target. I mean one time they went after a toilet paper company, I still don't understand what that toilet paper company did but it was an opportunity to make a splash. </s><s>[INST] Is there someone for the joke for the lulls. [/INST] It developed into that. So I think the hacktivism and the anonymous stuff wasn't so much for the locals. Um But from that H. B. Gary federal hack. Then there were six guys that worked well together and they formed a crew. A hacking crew and they kind of split off into their own private channels. And that was low Sekar or laughing at your security was their motto. </s><s>[INST] So that's L. U. L. Z. S. E. C. Lulls sec. [/INST] Of course it is. </s><s>[INST] SEK and who founded that organization. [/INST] So Kayla and Sabu were the hackers of the group and so they really did all the work on H. B. Gary. So yeah they're online names, their their their nicks. Um And so you know they they that's all they knew each other </s><s>[INST] as you [/INST] know, they talked as those names um And they worked well together and so they formed a hacking crew and that's when they started the at first they didn't name it. This was the 50 days of lulls where they would just release major major breaches. Um And it stirred up the media. I mean they put hacking in on in the media every day. Uh They had 400 or 500,000 twitter followers. Um You know, and it was kind of interesting. Um But then they started swinging at the beehive and they took out some FBI affiliated sites. Uh and then they started fuck FBI Fridays where every every friday they would release something and we waited it for bated breath. I mean they had a hook line and sinker pissed. We were waiting to see what was going to be dropped every friday. It was, it's a little embarrassing looking back on it now. </s><s>[INST] And this is in the early 2010s. [/INST] Yeah this was 2010, around there can </s> |
<s>[INST] driving cars. You know that, uh, from an outsider perspective, it's always frustrating to me. Well, I read a lot, some insider in a certain philosophical sense. Uh, the, it's frustrating to me how often trusts us used in surveys and how people say make claims that of any kind of finding they make about somebody clicking on answer. You just trust is uh, uh, yeah, behavior. Just you said it beautiful. I mean the action, your own behavior as is what trust is. I mean that everything else is not even close. It's almost like a absurd comedic, uh, poetry that you weave around your actual behavior So some people can say they, they trust, you know, I trust my wife, husband or not whatever. But the actions is what speaks volumes. You [/INST] bugged their car. You probably don't </s><s>[INST] trust them. I trust them. I'm just making sure. No, no, that's, yeah. [/INST] Like even if you think about cars, I think it's a beautiful case. I came here at some point. I'm sure on either Uber or lift, right. I remember when it first came out, right? I bet if they had had a survey, would you get in the car with a stranger and pay them? How many people do you think would have said like really? You know wait even worse would you get in the car with a stranger at one a.m. In the morning to have them drop you home as a single female? Like how many people would say uh That's stupid. And now look at where we are. I mean people put kids like right? Like oh yeah my my child has to go to school and I yeah I'm gonna put my kid in this car with a stranger. I mean it's just fascinating how like what we think we think is not necessarily matching our </s><s>[INST] behavior and certainly with robots for the town's vehicles and all the kinds of robots you work with that. It's uh yeah it's uh the way you answer it, especially if you've never interacted with that robot before. Uh If you haven't had the experience you're being able to respond correctly on the surveys uh Is impossible. What do you, what role does trust play in the interaction do you think? I guess it's good to is it good to trust a robot? What does over trust mean? Was it it's good to kind of how you feel about autopilot currently, which is like from a robotics perspective is like so very cautious. [/INST] Yeah. So this is still an open area of research but basically what I would like um in a perfect world Is that people trust the technology when it's working. 100% and people will be hypersensitive and identify when it's not but of course we're not there. Uh That's that's the ideal world and but we find is that people swing right? They tend to swing. Which means that if my first and like we have some papers like first impressions and everything is everything right. If my first instance with technology with robotics is positive um it mitigates any risk. It correlates with like best outcomes. Um It means that I'm more likely to either not see it when it makes the mistakes or false or I'm more likely to forgive it. Um And so This is a problem because technology is not 100% accurate. Right? It's 100% accurate although it may be </s><s>[INST] perfect. How do you get that first moment? Right. Do you think there's also an education about the capabilities and limitations of the system? Do you have a sense of how do you educate people correctly in that first interaction? [/INST] Again, this is this is an open ended problem. Um So one of the study that actually has given me some hope that I were trying to figure out how to put in robotics. So there was a uh a research study that has showed for medical ai systems giving um information to radiologists about, you know here you need to look at these areas um on on the X ray. What they found was that when the system provided one choice, there was this aspect of um either no trust or over trust. Right? Like I'm not going, I don't believe it at all or a um Yes, yes, yes. Yes. And and they would miss things right Instead when the system gave them multiple choices, like here are the three. Even if it knew like, you know, it had estimated that the top area you need to look at was he, you know, someplace on the X ray, if it gave like one plus others, the trust was maintained and the accuracy of the entire population increased. Right? So basically it was a you're still trusting the system, but you're also putting in a little bit of like your human expertise, like your human decision processing into the equation. So it helps to mitigate that over trust risk. </s><s>[INST] Yeah. So it's a fascinating balance of the strike. [/INST] I haven't figured out. </s><s>[INST] Again. It's open exciting open area research. Exactly. So what are some exciting applications of human robot interaction? You started a company? Maybe you can talk about the exciting efforts there, but in general also what other space can robots interact with humans and help? [/INST] Yeah. So besides health care, because you know, that's my bias lens, my other bias lenses. Education, I think that well when we definitely we in the US, you know, we're doing okay with teachers. But there's a lot of school districts that don't have enough teachers. If you think about the teacher student ratio for at least public education um in some districts, it's crazy. It's like how can you have learning in that classroom, Right, Because you just don't have the human capital. Um And so if you think about robotics, bringing that into classrooms as well as after school space, where they offset some of this lack of resources in certain communities, um I think that's a good place. And then turning on the other end is using these systems then for workforce retraining and uh dealing with some of the things that are going to come out later on of job loss. Like thinking about robots and ai systems for retraining and workforce development. I think that's exciting areas um that can be pushed even more and it would have a huge, huge impact. </s> |
<s>[INST] And that you know that that idea of motivation in general, we talked about goals and ai but goals aren't quite the same thing as like the our mortality. It feels like it feels like, first of all, humans don't have a goal and they just kind of create goals at different levels. They like make up goals because we're terrified by the mystery of the thing that gets us all. So we make these goals up. So we were like a gold generation machine as opposed to a machine which optimizes the trajectory towards a singular goal. So it feels like that's an important part of uh cognition, that whole mortality thing. [/INST] Well it is it is a part of human cognition but there is no uh reason for uh that mortality to come to the questions for a artificial system because we can copy that initial system. The problem with humans is that we can't clone you, I can I can I can even if I clone us uh you know the hardware, your experience that was stored in your brain, your episodic memory, all those will not be captured in the in the new clothes? Um So but that is not the same with an Ai system, Right? So but </s><s>[INST] it's also possible that the thing that you imagine with with us humans is actually funding of fundamental importance for intelligence. So like the fact that you can copy in their system means that that Ai system is not yet. Uh G I. So like there, so if you look at existence proof, if we reason based on existence proof, you could say that it doesn't feel like death is a fundamental property of an intelligence system, but we don't yet give me an example of an immortal intelligent being. We don't have those. It's very possible that, you know, that's that is a fundamental property of intelligence is a thing that has a deadline for [/INST] itself. You can think of it like this, I suppose you invent a way to freeze people for a long time. It's not dying, right? Uh So so you can be frozen and woken up thousands of years from now. Uh So it's no fear of death. So </s><s>[INST] we'll know the you're still it's not it's not about time, it's about the knowledge that it's temporary. Uh And the that aspect of it, the finite ness of it, I think um creates a kind of urgency, [/INST] correct for us, for humans, for humans. Yes. Uh And that that is part of our drives. Uh But um and that's why I'm not too worried about ai uh you know uh having motivations to kill all humans and those kinds of things. Why just wait, You know? So so why do you need to do that? I've </s><s>[INST] never heard that before. That's a good point. Yeah. Just murder. Seems like a lot of work. Just wait it out. Probably hurt themselves. Let me ask you. Uh people often kind of wonder world class researchers such as yourself. What kind of books technical fiction, philosophical were um uh had an impact on you in your life. And uh maybe ones you could probably possibly recommend that others read. Maybe if you have three books that pop in the mind. [/INST] Yeah, so I definitely liked judea Pearl's book, Probabilistic reasoning and intelligence systems. It's um it's a very deep technical book. But what I liked is that so there are many places where you can learn about probabilistic graphical models from. But throughout this book people kind of sprinkles his philosophical observations and and he thinks about can access to how the brain thinks and attentions and resources all those things. So so that whole thing makes it more interesting to read. </s><s>[INST] He emphasizes the importance of causality. [/INST] So that was in his later book. So this was the first book, probabilistic reasoning and intelligence systems. He mentions causality but he hadn't really sunk his teeth into like you know, how do you actually formalized? Yeah. And the second book Causality 2001. The one in 2000 that one is really hard. So I wouldn't recommend that. </s><s>[INST] Yes, so that looks at the like the mathematical, like his model of uh do calculus, calculus. Yeah it was pretty dense [/INST] mathematics. Uh the book of why is definitely more enjoyable for sure. Yeah. So yeah so I would I would recommend probably still listening in intelligent systems. Another book I liked was one from Doug Hofstadter. Uh This is a long time ago. He has a book, had a book I think called, it was called The Mind's Eye. It was uh uh probably have starter and Daniel Dennett together. Yeah </s><s>[INST] so and I actually was uh I bought that book on my shelf isn't ready yet but I I couldn't get an electronic version of it which is annoying because I read everything on Kindle. Okay uh you have to actually purchase the physical, it's like one of the only physical books I have because anyway there's a lot of people recommended it highly. So [/INST] Yeah and the 3rd 1 I would definitely recommend reading is um this is not a technical book, it is history, it's called, it's the name of the book. I think it's Bishop's Boys. It's about wright brothers and uh and their their their path and how it was. Uh there are multiple books on this topic and all of them are great. It's uh fascinating how flight was uh you know treated as an unsolvable problem and and and also you know what aspects did people emphasize? Uh, you know, people thought, oh, it is all about just powerful engines, you know, just need how powerful lightweight engines. Uh, and so, uh, you know, some people thought of it as how what can we just throw the thing, you know, through it? Catapults? Yeah, so it's a very fascinating and even after they made the invention, like people not believing it and the social aspect of the social aspect, it's different, you know, very, do </s> |
<s>[INST] So I've for for 5 6 years at M. IT really focused on this human side of driving question. And one of the big concerns is the micro sleeps drowsiness. These kinds of ideas. And one of the open questions was is it possible to computer vision to detect or any kind of sensors then I think about computer vision is you don't have to have direct contact to the person. Is it possible to detect increases in drowsiness? Is it possible to detect these kind of micro sleeps or actually just sleep in general? Among other things that distraction. These are all words that have so many meetings that so many debates like, like attention is as a whole, another one. Just because you're looking something doesn't mean you're loading in the information just because you're looking away doesn't mean your peripheral vision can't pick up the important information. There's so many complicated vision science things there. Um So I wonder if you could say something to uh you know, they say the eyes are the windows to the soul. Do you think um the eyes can reveal something about uh sleepiness uh through uh the computer vision, just looking at the video of the face. And Andrew Huberman and your friend have talked about this. I would love to work on this together. Should do. It's a fascinating problem. What drowsiness is a trick you want? So there's what kind of information there's a blinking and there's eye movement and those are the ones that can be picked up with computer vision. Do you think those are signals that could be used to say something about where we are in this continuum? [/INST] Yeah, I do. And I think there are a number of other features to I think um you know, aperture of I so in other words, partial closures, foreclosures, duration of those closures, duration of those partial closures of the eyelid. Um, I think there may be some information in the pupil as well because as we're transitioning between those states change. There are changes in what's called the automatic nervous system or technically it's called the autonomic nervous system part of which will control your pew pillory size. So I actually think that there is probably a wealth of information when you combine that probably with aspects of steering, angle, steering maneuver And if you can sense the pressure on the pedals as well. My guess is that there is some combinatorial feature that creates a 15 a type of you are starting to fall asleep and as the autonomous controls develop that it's time for them to kick in. Some manufacturers, auto manufacturers sort of have something beater version of maybe an alpha version of of this already starting to come online where they have a little camera in the wheel that I think tries to look at some features, </s><s>[INST] almost everybody doing this and it's very alpha. Uh, so, uh, you know, the thing that you currently have, some people have that in their car, there's a coffee cup or something that comes up that you might be sleeping. The primary signal that they're comfortable using is a steering wheel reversals. So, so basically using your interaction with the steering wheel and how much you're interacting with it as a sign of sleeping it. So if you have to constantly correct the car, that's a sign of like you starting to drift into micro sleep. I think that's a very, very crude signal. It's probably a powerful one. There's a whole another component to this which is, it seems like it's so driver and subject dependent. Uh the how our behavior changes as we get sleepy and drowsy seems to be different and complicated, fascinating ways where you can't just use one signal, it's kind of like what you're saying. There has to be a lot of different signals that you should then be able to combine. The hope is there's uh the searches for like Universal signals that are pretty damn good for like 90% of people. [/INST] But I don't think we need to take necessarily quite that approach. I think what we could do in some clever fashion is Using the individual. So what you and I are perhaps suggesting here is that there is an array of features that we know provide information that is sensitive to whether or not you're falling asleep at the wheel. Some of those let's say that there are 10 of them, you know, for me, seven of them are the cardinal features for you. However, you know, six of them and they're not all the same sort of overlapping are those for you? I think what we need is algorithms that can firstly understand when you are well slept. So let's say that people have sleep trackers at night and then your car integrates that information would be amazing, understands when you are well slept. And then you've got the data of the individual behavior unique to that individual snowflake, like when they are well slept, this is the signature of well rested driving. Then you can look at deviations from that and patent match it with the sleep history of that individual and then I don't need to find The sort of, you know, the one size fits all approach for 99% of the people. I can create very bespoke tailor like set of features, the Saville row suit of sleepiness features, you know, that would be like if you want to ask me about moonshots and crazy ideas, that's where I go. But to start with, I think your approach is is a great one. Let's find something that covers 99% of the people because the worrying thing about micro sleeps of course, unlike, you know, drugs or alcohol, which you know certainly is a terrible thing to be behind the wheel with those often you you react too late and that's the reason you get into an accident when you fall asleep behind the wheel, you don't react at all. At that point, there is a two ton missile driving down the street and no one's in control. That's why those accidents can often be more dangerous. </s> |
<s>[INST] Is there some aspect that you worry about? Where with a large audience there seems to be a certain aspect of human nature where people like like to see others destroyed. Uh do you worry about hurting people that don't deserve it? Or rather sort of attacking people there Grifter light, but they get like a giant storm of negativity towards them and therefore sort of overpowering lee, cancel them or [/INST] hurt </s><s>[INST] them, disproportionately [/INST] sure, I mean I try to be sensitive to my platform and as I've grown, I've tried to make sure my video topics have grown with me and like It does reach this tricky point where if you're exposing a grifter with like 50,000 subs, who's doing some harm? Are you punching </s><s>[INST] down? [/INST] And so far there's been enough high profile things that I can distract myself with, where this has never been a problem, You don't ever want to be um sort of like Sir Lancelot in retirement? Um where have you heard this analogy? Okay, so there's this great analogy where it's like Sir Lancelot is the guy who slays the dragon, right? He gets a lot of fame and he gets a lot of fortune for saving the dragon or at least a lot of, you know, people love him. But what happens after he slays the first dragon, </s><s>[INST] he's gotta go find [/INST] a bigger dragon. So he goes, finds a bigger dragon and eventually, depending on how many dragons you think are there in the world, maybe he kills all the dragons. And one day people go see Sir Lancelot and he's in a field with cows and he's chopping their </s><s>[INST] heads [/INST] and he's sort of put himself in retirement but he can't even enjoy the fruits because his whole thing is like I'm killing the dragon. So I try to be cognizant and I try to always make myself willing to hang up my, </s><s>[INST] my suspenders [/INST] I guess hang up my hat. I try to be aware like if I significantly improve the problem, I put myself out of business, I want to be okay with that basically. Um and just be fine with. I don't, the funny thing is I was more worried about this as like an issue earlier because I thought there was a finite, like I was like, I'm gonna solve this faster especially as it started gaining traction. Like I'm gonna solve this fast. I got this, you know, classic naive um you know, we all think we're so influential. Well yeah, and </s><s>[INST] you just, you [/INST] just, yeah, you just get like uh with time you get humbled because you talk to people I've talked to like versions of coffee zilla that are older and it's like, it's like oh yeah they didn't solve it and they probably were better. I just imagine </s><s>[INST] a smoke filled room [/INST] like </s><s>[INST] just like retired batman and you read this young bright eyed [/INST] oh yeah, very </s><s>[INST] spirited investigator, [/INST] exactly </s><s>[INST] what what's the process of investigation that you can speak to, what are some interesting um things you've learned about what it takes to do great investigations, sure [/INST] great investigations reveal something new or bring something to light. So I think what everyone thinks in terms of investigations is like a lot of like you know googling or like searching through articles, I think that's the first thing you want to get away from and you want to try to talk to people doing like the non obvious things and just trying to get perspectives that are beyond just what is available. So a lot of it's just having conversations is so enlightening um both the victims and also obviously trying to talk to the people themselves secondly, there's sometimes some analysis you can generate that's meaningful like Blockchain evidence. So in the case of Safe moon for example, going back to that I found someone's secret account where they were pumping dumping coins, they were saying things like who sold, I'll, you know, I'm so mad at the guy who sold f the guy who sold and you look at his account, he was the guy selling and it's like that is just, that's great stuff. So, digging through the Blockchain, kind of, I've gained some skills there. Um And that's kind of this fun, I guess, I would say it's this weird edge I have right now, because a lot of people don't know too much about that. Um and so I have this weird expertise that works now. I don't think that'll work forever because I think people kind of figure out how to do very similar analyses, but uh so it's it's like kind of an interesting edge right now that I have, </s><s>[INST] so that's like a data driven investigation, but you also do interviews, right? [/INST] Yeah, definitely. And then also recently I've tried to get more response speaking to your point about like as your platform gets bigger, you need more responsibility. I've tried to get much more um responsible about like reaching out or somehow giving the subject some way to talk, because I think in early on, I was such a small channel that ai if I asked them, they wouldn't answer. But b I kind of felt like I was launching these videos into the abyss, and when some of my videos had real traction, I was like, okay, hang on a second, let's double check this, let's triple check it. Let's try to make sure um all this stuff is correct and there's no other side of the story. I'll say this has interesting implications because for example, I investigated this thing called genesis their billion dollar crypto lender. And my conclusion was that they were insolvent. That's a huge accusation. So what do you do while I emailed their press team? Everybody, I said, hey, I think you're insolvent. I think you're this. I think I laid out all my accusations and I said you have till I think two p.m. The next day to respond at eight a.m. Before I made my video, they announced to all their investors that they're freezing withdrawals. They don't have the money. So they front, I don't know if they saw the, like, I don't know if they actually saw that even I don't want to take credit for collapsing them or whatever. But my point is, had I not taken that level of kind of care and just said, hey, you're, you're a scammer, you're frauds, ironically, could I have done more good by allowing people to withdraw their money early. I made some tweets that people did see that like some people got their money out, but my Youtube audience is much larger and could I have helped more people. Had I not given them basically the ability to know what I was going to produce when I produced it boy, </s> |
<s>[INST] Do you see, do you think for you violence would be compelling? [/INST] No, first of all I was too small. Um but second I just </s><s>[INST] dynamite doesn't care about your size. [/INST] Yeah, but I mean retribution does and I think I I don't know. But to me, violence is the kind of thing where you think you're running it, but it's running you once you, you know, cross that line, you know, violence sings its own song. So whenever I hear even contemporary times or people are advocating for, you know, violent actions, it's like when you start a fire, you're not like I'm just gonna burn down this house. You know, it's it's it's and there's many cases over and over of people who are building bombs or trying to assassinate someone or things like that and it ended up literally literally literally blowing up in their own face. So and violence doesn't really work necessarily because you know, if you have an assassination, you're not assassinating the presidency, you know, if you take out, you know, a president, there's another president instantly there. So what have you accomplished? Someone's husband, dad is gone, you replaced him with someone who's now is in a position to crack down and retaliate with even more violence. So it's it the calculus for me isn't there, would I be advocating for then, who knows? Um but I mean, I don't know if I'd be about this space to be, I certainly wouldn't have the space to be a podcaster or like a media personality that wasn't really a thing to some extent, it was in the 19 twenties with the Algonquin round table and all the people from the new yorker magazine. Um but they were all drunks, you know, it was very much um </s><s>[INST] a a [/INST] weird kind of situation to be a thinker. </s><s>[INST] What what would you think you do work at a carnival? You look good in lipstick, So [/INST] thank you. Um I look at anything. Um what would I I don't know, I mean you're not building robots. I mean you could have been a Tesla </s><s>[INST] right? I didn't [/INST] mean a car. I meant the person like </s><s>[INST] I understand, Oh thank you for explaining the witty comments to me. I [/INST] didn't at all because you wouldn't do an Einstein because he was an immigrant. </s><s>[INST] So I wouldn't work with an immigrant, what does that even mean? [/INST] You would have been a Tesla like figure. There's already a Tesla. So you wouldn't literally be Tesla. So I said a Tesla. </s><s>[INST] Oh Tesla. Okay, so alright, I thank you for the explanation. See Mike Michael doesn't only make funny things. He also explained [/INST] to them for you. It wasn't </s><s>[INST] Manse place them, it [/INST] wasn't funny at all </s><s>[INST] that I agree with. Okay, so yes, when, when you achieve this [/INST] is why Kanye didn't like you. It's this </s><s>[INST] Alright, I'm downgrading it from a nine down 28 if you keep talking like this. Um five is a real possibility. Alright, so vacuum is the kind of vacuum is created with violence is is usually um filled with like a with a harsh, with a harsher figure. So you don't think violent revolution ultimately leads to a positive to positive progress in the short term. [/INST] Well sometimes it does. The american revolution I think was a positive example and overthrowing the czar which was done peacefully uh was a positive example. But again, when violence happens, people get scared and they want the violence stopped immediately. And that's a call for authoritarianism. And you see it time and time again. And and they also want retribution. They were like bring this back to normal. Uh and they don't really worry about things like civil liberties or things like that. It's it's it's a very uh and then it also creates this space for invasion from foreign sources or demagogues. You know, like, oh look, they're killing us in the streets now. You got to support me. It's it's a very um deadly game. Obviously. </s><s>[INST] Remember somebody told me that uh I forget where it was, but they told me that from the very beginning, it was obvious that communism is an evil system that would, or a system that leads to evil. And uh to me at least that's not if I had to put put myself in the beginning of the 20th century at the end of the 19th century, that's totally not obvious. They are trying to elevate humanity. The basic worth of a human being, of a hard working human being of the working class of the people that are doing the work and striving and just uh really trying to build up society with their own hands. It seems like a beautiful ideal. Uh So I guess the question is, can you see yourself believing that in the ideas of socialism and communism? Yeah, let's say if you were living in Russia? [/INST] Oh yeah, easily. So, first of all, I don't think anything is obvious in politics. It's not obvious that, you know, uh humans have rights. It's not obvious that liberty is better or the markets either either whether you're for, you know, welfare state or your for more free markets, not those is obvious. Both of them involve an enormous amount of thought and background information. So when someone says something is obvious in politics, they really mean something is apparent. Well, it's not apparent on its face that if we all get together and promote a society based on equality and we all chip in that it's going to really be good for everyone. I mean, that to me is the promise of communism. Um and it was also very appealing to many people because it was new. So the idea was, alright, we've tried it these other ways. There's all these negative consequences, you have all these slums, you have people getting, you know fired and then they have no recourse you have women with 10 kids and they can't feed their kids infant mortality. You have, don't have sanitation, you don't have food, you know, everyone's illiterate and uneducated and then here saying, look, if we all chip in together, everyone will have close, everyone have food, everyone be educated, everyone will do their part. It's gonna be rough in the short period. That's a very compelling case to be made for communism. It's really easy in many ways when something hasn't been tried to make it sound uh, compelling because you just talk about how great it's gonna be. And then no one, no one, you know, people always arguing about like Venezuela and Sweden, like, you know, you want democratic socialism to be like Sweden, you don't want to be like Venezuela, the Venezuelans didn't vote for Venezuela, They voted for Sweden. They ended up with Venezuela. So it's, I think, and the thing with the communism, especially at that era, it was very much a correlated with people who are too smart for their own </s> |
<s>[INST] you're the one triggered person. He uh maybe he trolled the troll. [/INST] Well it worked. He yeah he did. The sharks lay eggs. </s><s>[INST] I'd like to get an updated 2021 version of Michael Morell was giving advice to young people. Okay, So there's God forbid high school students, college students listening to you uh and looking to you for advice. What advice would you give them about career and about life. How to live a life that can be [/INST] proud of. This happens a lot because I have my local community malice that locals dot com and there's a lot of young people in there. So that's a great place to give them a metal piece of advice. Don't ask your friends for advice because you're an idiot at your age and they're all idiots and they don't want to seem like idiots so there's going to give you advice, They pulled it from the tv and no one knows what you're talking about and it's just gonna be counterintuitive. So seek out advice from people who you seek to emulate them and ask them for advice if you can't get a hold of them figure out a way to get a hold of them incentivize them in some way. Um You'd be surprised how many people are responsive on twitter or in social media if you just ask them a basic life question because then they can quote tweeted an answer to the whole population so that would be one mechanism. Um It's also very hard at that age to realize your parents might not be all that bright and that might not be all that good people. Um So that's a hard one at that age to kind of wrap your head around just because they love you, doesn't mean they understand you and that's okay, that's that's okay, like everybody </s><s>[INST] uh your trump is pretty good too. I like your trump to talk to Ellen conversation. [/INST] Well Mr President, you know look uh something's you did like some not so much but you know for the most part I think they're kind of good thing, what are you talking about? Um hey guys what are you guys, what are we talking about, like I sucked up the lexx anyway so that those would be to piece the other piece of advice I would say is join the gym or have some kind of quantifiable daily improvement to keep you sane. So the reason I always say weightlifting and it could be uh running could be jump rope, I don't care what it is because if you have those numbers moving in the positive direction, psychological, if you're dealing with depression or anxiety, it's concrete proof to shut your brain up because your brain knows how to talk to you, your brain is off on your enemy. And I'll say exactly the right thing to undermine you. So that's an issue. Um, B I just, this works for me. Maybe it weren't for most people. I'm very high on the openness metric. Try new experiences and new things, try things you don't like. Um it's okay to have a bad experience. You've learned something. So go to a restaurant of a cuisine you wouldn't like or hadn't heard of read a book that's popular, but you have no interest in um read a lot. For example, I didn't know anything about the election. What was in 18 92 when there was like a split between the electors, read a book about it. Oh, I don't know anything. You know, I know anything really about Malcolm X. Read a book about him, you'll be amazed how much more fool you become as a person. </s><s>[INST] You see value in writing also like writing down your ideas. [/INST] No, I think there's very little value in that. I'm not being, I'm not joking because you're probably not going to revisit what you've written down. Um, but the </s><s>[INST] act of writing, you don't, you don't see it solidifies somehow thoughts in your mind. Not for me, it doesn't feel [/INST] like a tweet. Well, because then I have to have it narrowed down into like a phrase or the </s><s>[INST] responsibility of there being an audience. [/INST] No, I just meant in terms of I've got to 80 characters instead of having a brandy mandarin. Thought I have to codify it in something that's catchy in short. That's a good useful mental exercise. </s><s>[INST] What face do you make when you tweet? [/INST] I wouldn't know. I don't know. That's a good point. </s><s>[INST] I sit on the toilet. How much what percentage is on the toilet? [/INST] Very little on the toilets. I usually more reading. Um, so even though my tweets are all literally shit, very few of them are on the toilet. Um that's some advice. Um, don't compare yourself to other people. That's a really dangerous one. All my friends are married, I should have, I should have a kid by should there's an expression in recovery. Stop shooting yourself, but it's, but it should should should, it's it's stupid. I also, and this could be my hoarder brain, I surround my house with talismans of joy. So if you have an accomplishment, like when I did Rogan once I bought with the sock store and I bought these orange socks with black cherries on them. And now whenever I wore that socks, those socks, I'm like, oh, this is because I was on Rogan, that was kind of a big deal. So if you have these little things throughout your house, it's, it was good mental fuel. Even like, like a toy, Remember when I was a kid? Oh, you know what, this little moments that inspire happiness, I think are visually very useful. Um, so that's another one. Um, and I, </s> |
<s>[INST] Metal compression consciousnesses metal compression. [/INST] That's that's an eye and an idea and in some sense, you know the crazy. </s><s>[INST] Thank you. So uh but do you think if we think of touring machine, a universal turing machine, can that achieve consciousness? So is there some think beyond our traditional definition of computation that's required? [/INST] So it's a specific computation and I said this computation has to do with compression and the compression itself may be another way of putting it is like you are internally creating the model of reality in order like a like a you try insight to simplify reality in order to predict what's going to happen. And that also feels somewhat similar to how I think actually about my own conscious experience. So clearly I don't have access to reality. The only access to reality is through you know cable going to my brain and my brain is creating a simulation of reality and I have access to the simulation of reality. </s><s>[INST] Are you by any chance aware of? Uh the Hunter prize Marcus Hutter? He made this prize for compression of Wikipedia pages and uh there's a few qualities to it. One I think has to be perfect compression which makes I think that little quirk makes it much less um applicable to the general task of intelligence because it feels like intelligence is always going to be messy. Uh like perfect compression is feels like is not the right goal but nevertheless a very interesting goal. So for him intelligence equals compression. And so the smaller you make the file given a large Wikipedia page, the more intelligent the system has to be. [/INST] And that makes sense. So you can make perfect compression if you start errors. And I think that actually what he meant is you have algorithm plus errors by the way, who hotter is uh he was uh gave the advisor of Shane Black who is defined uh Deepmind co founder </s><s>[INST] Yeah, yeah, so there's an interesting uh and now he's a deepmind, There's an interesting uh network of people. He's one of the people that I think seriously took on the task of what would an A G. I. System look like I think for the longest time the question of A G. Was not taken seriously or rather rigorously. And he he did just that like mathematically speaking, what would the model look like if you remove the constraints of it having to be uh um having to have a reasonable amount of memory, reasonable amount of running time, complexity uh computation time, what would it look like? And essentially it's it's a half math, half philosophical discussion of how would like a reinforcement learning type of framework look like from a [/INST] gi Yeah. So he developed a framework even to describe what's optimal with respect to reinforcement learning. Like there is a theoretical framework which is as you said under assumption there is infinite amount of memory and compute and there was actually one person before his name is Solomon of who their extended uh Solomon of work to reinforcement learning but there exists uh theoretical algorithm which is optimal algorithm to build intelligence. And I can actually explain to the </s><s>[INST] algorithm. [/INST] Let's </s><s>[INST] go, let's go. [/INST] So the task itself, </s><s>[INST] can I just pause how absurd it is for brain in a skull trying to explain the algorithm for intelligence. Just go ahead. [/INST] It is pretty crazy. It is pretty crazy that you know, the brain itself is actually so small and it can ponder uh </s><s>[INST] how design algorithms that optimally solve the problem of intelligence. Okay. All right. So what's the algorithm? [/INST] So let's see. So first of all the task itself is described as you have infinite sequence of zeros and ones. Okay? You read and beats and you are about to predict and plus one bit. So that's the task. And you could imagine that every task could be casted as such a task. So if for instance you have images and labels you can just turn every image into a sequence of zeros and ones then label you concatenate labels and you and that that's actually the the and you could you could start by having training data first and then afterwards you have test data. So theoretically any problem could be casted as the problem of predicting zeros and ones on this infinite tape. So um so let's say you read already end bits and you want to predict and plus one bit and I will ask him to write every possible program that generates these and bits. Okay. So um and you can have you choose programming language, it can be packed on R. C. Plus plus and the difference between programming languages might be, there is a difference by constant A symptomatically your predictions will be equivalent. So you read and beats you enumerate all the programs that produce this and and beats in their output and then in order to predict And plus one bit you actually wait. The programs according to their length And there is like some specific formula how you weight them. And then the end plus one bit prediction is the prediction uh from each of this program according to the weight </s><s>[INST] like statistically statistically pick. So the smaller the program the more likely you are to pick the its output. So uh that's that algorithm is grounded in the hope Or the intuition that the simple answer is the right one. [/INST] It's a formalization of it. Um It also means like if you would ask the question after how many years would you know sun explode? You can say it's more likely dancer is to to to some power because the shorter program. Yeah. And I </s> |
<s>[INST] That's beautifully. That's so beautifully put and so true. Yeah, it's Mhm. Okay. It's worth pausing on that this dismissal, people forget the beauty of that curiosity. That's true in the scientific field too, is uh uh that that youthful joy of like everything is possible and we can understand it with the tools of our mind. [/INST] Yes. And that's what it's all about. That's what I mean. His ideas at the end of the day all bow down to is that confidence and that passion and that curiosity and that interest. And if you, you know, think about what academia does to so many of us, right? We go into academia and and we're excited about we're gonna learn stuff, we're going to discover things and then they stick you into sub sub field and examining some minutia that's insignificant and unimportant. And and to get published, you have to be conventional, you have to do what everybody else does. And then there's the tenure process of seven years where they put you through this torture, the right papers that fit into a certain mold and by the time you're done, you're in your mid thirties and you've done nothing, You've discovered nothing. You you you all in this minutia in this stuff and it's destructive and where's holding onto that passion, holding on to that knowledge and that confidence is hard and when people do away with it to become cynical and they become part of the system and they inflict the same pain on the next guy that they suffered because that's part of how it works. </s><s>[INST] Yeah, there's uh, this happens in artificial intelligence. This happens when like a young person shows up and with like fire in their eyes and they say, I want to understand the nature of intelligence and everybody rolls their eyes well for these same reasons because they've spent so many years and a very specific set of questions that that kind of they compete over and the right papers over and have conferences about. And it's true those that incremental research is the way you make progress answering the question of what is intelligence exceptionally difficult. But when you market you actually destroy the reality is when we look like centuries from now, we'll look back at this time for this particular field of artificial intelligence, it will be the people who will be remembered will be the people who have asked the question and made it their life journey of what his intelligence and actually had the chance to succeed most will fail asking that question, but the ones that had a chance of succeeding and I had that throughout their whole life. Uh, and I suppose the same is true for philosophy. It's in [/INST] every field. It's it's it's asking the big questions and staying curious and staying passionate and staying excited and accepting failure. Right? Accepting that you're not gonna get the first time, you're not gonna get the whole thing, but, and and sometimes you have to do that, my new show work and I'm not here to say nobody should specialize, you shouldn't do that. I know you have to do that, but there has to be a way to do that work and keep the passion and keep and keep it all integrated. That's another thing. I mean, we don't live in a culture that integrates right? We live in a culture that is all that is all about, you know, this minutia and not, and you know, medicine is another field where you you specialize in the kidney. I mean, the kidneys connected to other things you've got to and we don't have a holistic view of these things. And I'm sure an artificial intelligence, you're not going to make the big leaps forward without a holistic view of what it is you're trying to achieve. And maybe that's the question of what is intelligence, but that's the kind of questions you have to ask to make big leaps forward to really move the field in a in a positive direction and it's the people who can think that way, who move fields and moved technology. You move anything anything is is everything is </s><s>[INST] Just like you said, it's painful because underlying that kind of question, is well, maybe the work I've done for the past 20 years was it was a dead end and you have to kind of face that even just, it might not be true, but even just facing that reality, it's just it's a it's a painful feeling. Absolutely. But [/INST] it's that's part of the reason why it's a point to enjoy the work that you do. So even if it doesn't completely worked out during the process right? It was not a waste because you enjoyed the process. And if you learn as any entrepreneur knows this right? And if you learn from the waste of time from the errors from the mistakes, then you can build on them and make things even better. Right? And so the next one years, uh massive success. </s><s>[INST] Can we uh another impossible task. So you did wonderful Ian talking about Iran the other impossible task of giving a world when overview of the philosophy of objectivism, the philosophy of Ayn rand. [/INST] Yeah. So luckily she did it in an essay, she talks about doing your philosophy on one ft. Um but let me integrate it with the literature and with her life a little bit. She wanted to be a writer, but her goal, she had a particular goal in her writing. She was an idealist, right? She wanted to portray the ideal man. So one of the things you do when you want to do is what is an ideal man, You have to ask that question. What does that mean? You might have a sense of it, you might have seen glimpses it glimpses of it in other people's literature. But what is it? So she starts reading philosophy to try to figure out what philosophers say about the ideal man in which he finds horrifies her in terms of the view of most philosophers of man. And and she's she's attracted certainly when she's young teen each year, because nature at least has a vision of of of grandeur for man, even though his philosophy is very flawed and has other problems and contradicting man in many ways, but at least he has that vision of what is possible too man. And she's attracted to that romantic vision, that idealistic vision. So she discovers in writing and particularly in writing out the shrug. But even in the fountain that she's going to have to develop her own philosophy, she's going to have to discover these ideas for herself because they're not fully articulated anywhere else. The glimpses again of it in Aristotle in in in each but they're not fully fleshed out. So, to a large extent, she develops a philosophy for a very practical purpose to write, to write a novel about the ideal man. And and and atlas shrugged is the manifestation of that, </s> |
<s>[INST] Um, [/INST] so at some point, you know, he never knew a lot of openings in chess, but I taught him, I taught him a couple of openings as black. Um it's the, It's the special Sicilian that I've played a lot myself also. Um during the World Championship in 2018, I won a bunch of games in 2019 as well, so that's one opening and also taught him as black to play the, in defense. Um, and then so the in defense goes like it goes like this, um, it's characterized by, by this bishop move. Um, and so he would play those openings pretty pretty exclusively and as black and the tournaments that it did play and also the special of Sicilian is like, that's the only two of my sisters play have played a bunch of chess tournaments as well and that's the only opening they know as well. So my, my family's reporter is very narrow so so this is the, this is the system, black goes here and then very often white takes a pawn and black takes the pawn. Um, so at some point I was watching one of my, my father's online blitz game this game and as white, he played this, this, so this is called the car can defense, he took upon. Um, it was taken back then he went with a night, um, its opponent one here and then you play the bishop here. So I've never seen this opening before. I was like, wow, how on earth did he come up with that and he said no, I just played the with the different colors because if the night was here it would be the same position. I was like, I never, I was like how am I like one of the best 20 players in the world and I've never thought about that. So I actually started playing, I started playing this line uh as white with pretty decent results and results and it actually became kind of popular and everybody who asked about the line, like I would always tell them, yeah, that's the Henry Carlson, </s><s>[INST] I [/INST] wouldn't necessarily explain why it was called that I would just always call it that. So I really hope at this point at some point this line will be uh </s><s>[INST] will find [/INST] its rightful name </s><s>[INST] in. Yeah, finds its way into the history books. Can you, what, what, what did you learn about life from your dad? What what role has your dad played in your life? [/INST] He's taught me a lot of things, but most of all, as long as you win a chess, then everything else is fine. Um I think my, especially my father, but my parents in general, they, they always wanted me to get a good education and find a job and so on. Um even though my father loves chest and he wanted me to play chess. I don't think he had any plans for me to be professional. I think things changed at some point, Like </s><s>[INST] I [/INST] was less and less interested in school and for a long time we were kind of going back and forth fighting about that, especially my father, but also my mother a little bit, it was at times a little bit difficult. </s><s>[INST] Yeah, [/INST] they sort of wanted me to do more school to have more options and then I think at some point they just gave up. But I think that sort of coincided when I was actually starting to make real money of tournaments and after that, you know, everything's been sort of easy and like in terms of the family, like they've never </s><s>[INST] put [/INST] any pressure on me or they've never put any demands on me. Um, they're just, yeah minus has to focus on chess. That's that, that's um, that's it. Like I think they taught me in general to be curious about the world and to get a decent general education, not necessarily from, from school, but like just knowing um, about the world around you and knowing history and being, you know, just being interested in society. I think in that sense they've done well </s><s>[INST] and he's been with you throughout your chest career. I mean there's something to be said about just family support and love that you have that, that, you know, this world is a lonely place. It's good to have people around you, they're like uh they got your back kind of you [/INST] know uh basically shape but I think to some extent all the people you surround yourself with they can help you a lot. It's the only family that only has their own interests at heart. And so for that reason like my father is like the only one that's been like constantly in the team that and that has always been around and it's it's for that reason that I know he has my back no matter what </s><s>[INST] now there's a cliche question here but let's try to actually get to some deep truth perhaps. But people who don't know much about chess seem to like to use chess as a metaphor for everything in life. But there is some aspect to the decision making to the kind of reason and involved in chess that's transferable to other things. Can you can you speak to that in your in your own life and in general the kind of reasoning involved with chess, how much is that this transfer to life out there? [/INST] It just helps you make decisions of </s><s>[INST] all of all kinds. [/INST] That would be my main takeaway that you learn to make informed guesses in a limited amount of time. </s> |
<s>[INST] I know how that goes. But she [/INST] tried but she said to him, but even the dog is allowed to eat the crumbs that fall from the master's table and he went, helped, helped. Now let's go back to just said about Putin or if you don't quote alone and myself and my family, of course the family is first, but once the family is good, it has to then spread to the community dan to the ST country world. The problem we have sometimes is that, and this is the reason why a lot of powerful families was overthrown. Like why do they had their own king with the guillotine? Because that once the family was strong, they didn't let the the wealth, the opportunity expand out. You know, look at the wu tang, Yes, our family was made strong first. But then all the woman was able to form their own corporations and they had their own sub families, it has to grow </s><s>[INST] out and they took over the world. You've talked about being vegan. I don't think I heard you explain this because it connects somehow about how you think about life. So you talk about when your family is good, you grow that like circle of empathy. You grow the community. Is that how you think about being vegan? That just the capacity of living beings on earth to suffer, that you just don't want to add suffering to them? [/INST] Yeah, I mean you said it clear, it's like nothing in all reality. I came to a realization and nothing really has to die for me to live. No animal, the plants themselves. Right? So this is say, you know, you want a steak, which is probably the most, you know, you know, I don't know the most expensive piece of me, but let's just say the steak is, you know, top of the line, nice steak and you're eating a steak for the protein to help build your muscle And I don't know if you got it from a cow or bull, but whether it's a cow or the ball, they grow to about £1,500 and it was a board, it's all mostly muscle and it's only eating grass. Uh huh. </s><s>[INST] Yeah. Yeah, there's there's yeah, it's possible to uh both as an athlete and just as a human being to perform well without meeting me. There's something, especially in the way we're treating animals to deliver that meat to the plate. I think about that a lot. So I I do I'm a robotics person ai person and I think a lot about, I don't know if you think about this kind of stuff but building A. I. Systems as they become more and more human. Like you start to ask the question of are we okay? Uh if we give the capacity for AI systems to suffer first to feel, but then to suffer um to hate and to love to feel emotion, how do we deal with that? It starts asking the same questions you ask of animals. Are we, are we okay adding that suffering to the world? [/INST] And I don't think we should have the suffering because it's not necessary? Like look if it's necessary, right? Because we're, you know, survival of the first law of nature of self preservation if you are in the desert and there's nothing else to eat but that lizard. Yeah. Okay. You gotta do what you gotta do. This has gotta go, you gotta go, you gotta do what you gotta do because at the end of the day, man, it's when they say man has Dominion over these things. His Dominion is almost like a caretaker how the way we do our Dominion, we dominated eat it cook it. Like who's the first guy that looked at the lobster. It was like I'm gonna eat this thing like it's supposed to, it's hard to eat it, you gotta go through a process to get that crab. Remember we see crabs when we was kids and I didn't know why I was always getting itchy throats all that. You know, you care, you don't know just eat. But at the end of the day, a crab didn't provide no more than a finger work for meat, mate. And it was hell, getting that, Getting it out. It's not worth it. In all reality, you could have gave me a, you could have gave me a banana and did better for my body and my appetite. Am I being fulfilled as full? Look, look at the blessings of of of of life, right? If you take a seed, oh, you get an apple and you eat it in that apple is multiple. Season it if you plant that seed, it will give you a whole tree with a whole bunch of apples with all multiple seeds. But if you kill a fish, it can't be produced. It's done. If you kill a cat is done, it's not, it's not coming back. But when you deal with the plants, even after you eat the apple and then you defecate, you're defecation is what feeds the ground, the cause of apple to grow </s><s>[INST] more. Yeah. Circle of life. And especially there's a there's a guy named David Foster Wallace, he wrote a short story called Consider the lobster. If you actually think philosophically about what from a perspective the lobster that's like symbolic or something because you're basically put in the water like cold water and then it heats up slowly until until it's no [/INST] more. Then they started eating lobsters in the acquisition. Yeah. </s><s>[INST] Just enjoy. So they were probably enjoyed torturing animals and they realized they're also delicious after the torture is finished. That's probably how they discovered it. [/INST] Let me ask you a question. I know you're asking me the questions but I want to talk a little bit about the ai and you said something about trying to put the emotion in it. Yeah. Right. Um So do we so are you thinking there's an algorithm for emotion? Yes. </s> |
<s>[INST] Sure, [/INST] yeah, first of all, I much prefer Janz metaphor. Um Miles Davis's was a real explorer in jazz and um, he had a coherent story, so I think I have one and, but it's not just the one you lived, it's the one you think about later. What a good historian does is they look back and they revisit. Um, I think what's happening right now is not a i that was an intellectual aspiration that's still alive today as an aspiration. But I think this is akin to the development of chemical engineering from chemistry or electoral entering from, from electromagnetism. So if you go back to the thirties or forties, there wasn't yet chemical engineering, there was chemistry, there was fluid flow, there was mechanics and so on. Um, but people pretty clearly viewed uh, interesting goals to try to build factories that make chemicals products and do it viably safely make good ones do it at scale. So people started to try to do that of course. And some factories worked. Some didn't, you know. So we're not viable. Some exploded. But in parallel developed a whole field called Chemical engineering. Right? And chemical engineering in the field it's it's no no bones about it has theoretical aspects to it. It has practical aspects. It's not just engineering, quote unquote, it's the real thing. Real concepts were needed, you know? Same thing with electrical engineering. You know, there was Maxwell's equations which in some sense where everything you know about electromagnetism but you needed to figure how to build circuits, how to build modules, how to put them together, how to bring electricity from one point to another safely and so on so forth. The whole field to develop code electrical injury. All right. I think that's what's happening right now is that we have we have a proto field which was statistics computed more the theoretical side of algorithmic side of computer science that was enough to start to build things. But what things systems that bring value to human beings and use human data and mix in human decisions. The engineering side of that is all at hawk. That's what's emerging. In fact if you want to call machine learning a field, I think that's what it is that the proto form of engineering based on statistical and computational ideas of previous generations. </s><s>[INST] But do you think there's something deeper about ai in his dreams and aspirations. As compared to chemical engineering? Electrical engineering? [/INST] Well, the dreams and aspirations maybe. But those are from those are 500 years from now. I think that that's like the Greek sitting there and saying we need to get to the moon someday. Um I think we have no clue how the brain does computation. We just a clueless. We're like, we're even worse than the Greeks on most anything interesting. Uh Scientifically of our era. </s><s>[INST] Can you linger in there just for a moment because you stand not completely unique, but a little bit unique in the in the clarity of that. Can you, can you elaborate your intuition of why we're like where we stand in our understanding of the human brain. And a lot of people saying you're scientists say we're not very far in understanding human brain, but you're like you're saying we're in the dark here. [/INST] Well, I know I'm not unique. I don't even think in the clarity. But if you talk to real neuroscientists that really study real synapses are real neurons. They agree they agree it's 100 hundreds of your task and they're building it up slowly. Surely what the signal is? There is not clear. We think we have all of our metaphors. We think it's electrical maybe it's chemical. It's a it's a whole soup, it's ions and proteins and to sell and that's even around like a single synapse if you look at a electro micrografx of a single synapse? It's a it's a city of its own and that's one little thing on a dendritic tree which is extremely complicated electrochemical thing. And it's doing these spikes and voltages have been flying around and then proteins are taking that and taking it down into the D. N. A. And who knows what? Um So it is the problem of the next few centuries. It is fantastic. But we have our metaphors about it. Is it an economic devices? It like the immune system or is it like a layered, you know, set of comedy? You know er thematic computations what we've all these metaphors and they're fun. Um But that's not real science per se. There is neuroscience that's not neuroscience. All right. That that's that's like the greek speculating about how to get to the moon fun, right? And I think that I like to say it's this fairly strongly because I think a lot of young people think they were on the verge because a lot of people who don't talk about it clearly let it be understood that yes, we kind of this is the brain inspired were kind of close breakthroughs are on the horizon and that scrupulous people sometimes who need money for their labs. Um That's what I'm saying. Unscrupulous but people will over sell um I need money from them. I'm gonna I'm studying your computational neuroscience um I'm going to over sell it. And so there's been too much of that. </s> |
<s>[INST] Why do you think there's not more excitement slash money behind this idea? [/INST] Currently there was people view that project as a failure, I think that they Confused the failure of a specific instance that was conceived 30 years ago for the failure of an approach, which they don't do for deep learning. So, you know, in 2010 people have the same attitude towards deep learning. They're like this stuff doesn't really work. And you know, all these other algorithms work better and so forth. And then certain key technical advances were made. But mostly it was the advent of graphics processing units that change that. It wasn't even anything foundational in in the techniques and there was some new tricks, but um mostly it was just more compute and more data, things like Image net that didn't exist before. Um that allowed deep learning and it could be to work, it could be that, you know, psych just needs a few more things or something like like, But the widespread view is that that just doesn't work and people are reasoning from a single example, they don't do that with deep learning. They don't say nothing that existed in 2010 and there were many, many efforts in deep learning was really worth anything. Right. I mean really there's no model from 2010 in deep learning press, this is the Deep learning, there's any commercial value whatsoever at this point, right? There all failures. Um, but that doesn't mean that there wasn't anything there. I have a friend um I was getting to know him and he said he said, I I had a company to, I was talking about, I had a new company, he said I had a company to and it failed and I said, well what did you do? And he said Deep Learning And the problem was he did it in 1986 or something like that and we didn't have the tools then, or 1990. We don't have the tools then. Not the algorithms, His algorithms weren't that different from modern algorithms, but he didn't have the Gps to run it fast enough, he didn't have the data and so it failed. Um it could be that, you know, symbol manipulation per se with modern amounts of data and compute and maybe some advance in compute for that kind of computer um might be great. My my perspective on it is not that we want to resuscitate that stuff per se, but we want to borrow lessons from it, bring together with other things that we've </s><s>[INST] learned and you might have an image net moment where it will spark the world's imagination and there will be an explosion of symbol manipulation efforts. Yeah, I think [/INST] that people at ai to paul Allen's uh institute are trying to trying to build data sets that well, they're not doing it for quite the reason that you say, but they're trying to build datasets that at least spark interest in common sense reasoning to create benchmarks. That benchmarks for common sense. That's a large part of what the Ai too dot org is working on right now. </s><s>[INST] So speaking of compute Rich Sutton wrote a blog post titled bitter lesson. I don't know if you've read it, but He said that the biggest lesson that can be read from 70 years of Ai research, is that general methods that leverage computation are ultimately the most effective do you think most effective [/INST] at what? Right. So so they have been most effective for perceptual classification problems and for some reinforcement learning problems. He works on reinforcement. </s><s>[INST] Let me push back on that. You're actually absolutely right. But I would also say that would be most effective generally because everything we've done up to, what would you argue against that is to me deep learning is the first thing that has been successful at anything in AI. And you're pointing out that this success is very limited folks. But has there been something truly successful before deep learning? [/INST] Sure. I mean I mean I want to make a larger point but on the narrower point, you know, classical AI is used for example in doing navigation instructions. </s><s>[INST] You know, it's very [/INST] successful. Everybody on the planet uses it now like multiple times a day. That's a measure of success, right? So I mean I don't think classically I was wildly successful but there are cases like that there is used all the time, nobody even notices them because they're so pervasive. So there are some successes for classically, I think deep learning has been more successful. But my usual line about this and I didn't invent it, but I like it a lot. It's just because you can build a better ladder, it doesn't mean you can build a ladder to the moon. So the bitter lesson is if you have a perceptual classification problem, throwing a lot of data that it is better than anything else. But that has not given us any material progress in natural language, understanding common sense reasoning like a robot would need to navigate a home problems like that. There's no actual progress there. </s><s>[INST] So flip side of that. If we remove data from the picture, another bitter lesson is that you just have a very simple algorithm and you wait for compute to scale. This doesn't have to be learning. It doesn't have to be deep learning. It doesn't have to be data driven, but just wait for the compute. So my question for you, do you think compute can unlock some of the things with either deep learning or symbol manipulation? That. [/INST] Sure, but I'll put a proviso on that. More computers always better. Like nobody's going to argue with more computers like having more money. I mean, </s><s>[INST] you know, there's the returns are more money. [/INST] Exactly. There's diminishing returns on more money. But nobody's gonna argue if you want to give them more money, right? Except maybe the people who signed the giving pledge and some of them have a problem. They promised to give away more money than they are able to. Um but the rest of us, you know, if you wanna give me more money, more money more problems. But okay, that's true to what I would say to you is your brain uses 20 watts and it does a lot of things that deep learning doesn't do or or the symbol manipulation doesn't do that? That Ai just hasn't figured out how to do? So it's an existence proof that you don't need, you know, server resources that are Google scale in order to have an intelligence. You know, I built with a lot of help from my wife to intelligences that are 20 watts each and far exceed anything that anybody else um has built at a silicon. </s> |
<s>[INST] of Satoshi's writing to Sabo, Satoshi's writing to Sabo. So I meant telemetry as a field. I didn't actually look for application for this particular problem. So you're saying Nick Sabo is the closest match? [/INST] Yeah. Somebody did it years ago. I didn't look at if the model was sound or not, but I just remember reading on a Bitcoin talk and Sabo is another one of the common candidates. </s><s>[INST] How funny Sabo and Adam are probably the top three things that people could list. What do you think about this idea of anonymity, of publishing something anonymously? Would you ever consider publishing a paper? You've been part of, I mean, the Cardano ecosystem has published a lot of incredible papers. Is there ever a value to publish anonymously? [/INST] Well, every paper that goes through the referee process, the authors are ripped off. So you don't actually see the authorship when you submit to the conference. That's just best practice. But the question is, do you preserve the anonymity post-conference and actually not reveal the author of the paper? It's a detriment for the deals we make because the whole premise of working with our company as an academic is that you're going to have amazing co-authors and your work is going to appear in great conferences, great journals, and as a consequence, you get tenure. If you publish anonymously, it's like doing clearance work in high energy physics or something like that. So after 30 years of this amazing career working on nuclear weapons and classified reactors, you finish and then you go to apply for a job and they're like, so what have you done for the last 30 years? Stuff. Where is it on your CV? Well, I can't really talk about it. Okay, welcome to community college. So you get really screwed if you do that. So there's a misalignment of incentives in the academic world towards anonymity. And generally, it's only done when you're doing something very controversial or there's a whistle blowing type of a component. It's not typically done for foundational work. And Satoshi was really one of the first things, because if there was a Satoshi docs themselves, I don't think it's possible anymore. But if he or she did that, that's like a Nobel Prize in economics, likely. You're on the shortlist for that. There's enormous accolades that become beyond the monetary incentives of being able to dox yourself. </s><s>[INST] That'd be cool if they give a Nobel Prize in economics to an anonymous, to Satoshi Nakamoto. [/INST] It's been proposed. It's been proposed and it was turned down. Yeah. So, yeah, I mean, there are a few people in our company that have done sedonymous publications, like the, if you look at the chimeric ledgers paper, that's a, it's not a real name. It's a crazy name. It's a sedonymous publication and, you know, but that's usually for throwaway work. There is one project we inherited from an anonymous person, which is fascinating. It's called Qeditas. And it's basically an extension of the Qed manifesto from the nineties. And the pseudonym is Bill White. And I think it's some anonymous mathematician, but I can't figure out which one it is. But basically it's a marketplace for deduction. So it's like this magic machine where you can create incentives for people to write mathematical proofs in a theorem prover and make some money from it. So there's some cool work that's there. And it's sad that Bill stayed anonymous because I think that could have been easily published </s><s>[INST] and there was a lot of really cool things that could have been done with Qeditas. So, you did say the success of Cardano, the vision you have is for you to have less and less power over time. So this idea of governance, what's your vision for a decentralized secure governance system [/INST] over time? So the first thing you have to do is you've got to look at meaningful metrics, not vanity metrics. So, what does it mean to have legitimacy in a governance system? You can build any governance system you want. You have a dictator, right? Like Bob is in charge. It's like whether you like Bob or not, he's in charge. It's not very legitimate. And you can have pure democracy where every single person votes and then nothing ever gets done. County dog catcher is like a six year election or something like that. So there's a spectrum there between absolute power to one and perfectly egalitarian power to every single potential participant inside the system. And then the question is, okay, well, how do you even architecture? How do you handle choice architecture in that? So are you asking your people about every question or are you asking your people about a subset of questions related to a particular set of topics but then they're not allowed to talk about other topics? Like for example, are they allowed to change the tax rate but they can't change freedom of speech? That kind of a thing. So the first thing you have to do when you build these types of systems and they get to a certain scale is you have to build some mechanism for people who are interested in governance to self-select and participate. Just create a collection bucket. In Bitcoin, we had Bitcoin Talk and Bitcoin Reddit and these things, and eventually the GitHub repos and these things, there was a place to go if you were interested. And you need some sort of change management system where people who want to evolve the system can write it down in a very careful way. So in Bitcoin's case, it was Bitcoin Improvement Proposal and Theorem, it's the Ethereum Improvement Proposal. And for us, it's the SIP, the Cardano Improvement Proposal, but just a structured way of discussing how you wish to change. Then there's a question of do you want to do this implicitly or explicitly? It's a case of Bitcoin and Ethereum it an implicit system. So there's no on chain voting. There's no like five people said this and four people said this. So we do this, in the case of Cardano, we're actually explicitly inviting us. This is one of the biggest differentiators between Cardano, Polkadot, and these other things is that we're really serious about governance to the extent that we're actually doing foundational research in e-voting. We're building new voting systems. We're exploring preference voting and quadratic voting. And the long and the short is that we want more and more people to participate in voting for things. And you have to start somewhere. So our hypothesis is we can bootstrap the system with the treasury system. So in Cardano, some of the inflation goes to the block producers, just like any other cryptocurrency, but some goes into a decentralized treasury. China has over a billion dollars of ADA in it. So it's a lot of money. And that treasury is not under my control or the foundation's control. It's actually controlled by the community as a whole through a program called Catalyst. And so all these people can come together and they can submit spending ballots and other people who hold ADA can vote to approve that. What's nice about it is you have a growth engine to improve two accesses. One is absolute participation. So increase the amount of people who hold ADA participating. So the absolute number of people participating. And the other is meaningful participation. So the depth of participation. Did you just show up and vote? Or did you spend hours debating things on idea scale, the innovation management platform, interacting with the funding proposal, going back and forth. And there's dozens of little things like that. Now my hypothesis is if you run this with enough iterations, eventually you get to a certain critical mass, like over 50% absolute participation and a high level of meaningful participation, where you can move beyond funding and you can start actually having meaningful questions about protocol design and improvement proposals and so forth. And then what you can do is you can roll out new voting systems and new social structures. And you can let them start voting on training wheels like system parameters. Like for example, the minimum transaction fee or that K parameter for the amount of stake pools or these types of things, okay? Then they build enough competency there and they move up a next level and they actually start talking about hardforks using the update system, the hardcore combinator system. See, so that's how you vote on a hard fork, voting on a hard fork. You can't do it unless your social dynamics are right and your voting system is right. And by the way, you also need to write a constitution around the same time, because not all hard forks are created equally. The kinds of things that would add support for a new cryptographic primitive are distinctly different from the kinds of things </s> |
<s>[INST] Yeah, that's kinda what that thought thought was. That wasn't the best land. Back in my day, we didn't have general relativity. Exactly, yeah. Otherizing, yeah. Yeah, that's kinda the point. So for the insults you have from the white guy to the white guy, the insults are much harsher. So when you start to apply [/INST] the same kind of harshness to other groups. You can make them feel like they really don't belong, and that otherizing effect is something is very hard for me. I can't really emphasize it because I never felt it. So I have to intellectualize it and then sympathize with it. It's like a whole process I have to go through. And then I try to walk other people through that because if you're a white guy on the internet, which is a lot of the internet, you really don't know what that feels like. You've never felt like that before. </s><s>[INST] You never- Well so you're now in a leadership position, grandpa destiny, so that's a lot of people who come to you for that. For that sort of pathway to empathy. Yeah. Not to other eyes. I mean, you have felt other eyes You mentioned the high school people not being- [/INST] Yeah, but those were always for things that like. It's different to insult somebody for a non-immutable characteristic. Like, okay, you think poorly about me because I'm not enough money or what I don't have money, but I could get more money and I could change that. But it's different for somebody to really attack you </s><s>[INST] for your gender or attack you for like your race. A lot of the attacks that hit the hardest is not about gender. I do think that the way women are attacked on the internet. It's the same kind of attacks [/INST] you would do towards other guys, but you'd go harsher. I feel like they're fundamentally different. I feel like when we're attacking guys, I'm not usually attacking you on the virtue of you being a guy. But if it's a woman and she's typing something like, oh, did your boyfriend type that for you? Or what are you even doing here? Shouldn't you be trying to find a husband? Or oh, you're a stupid kind of like, go start an OnlyFans or whatever. No, but the stupidity, the intelligence aspect is what's attacked. Yeah, but it's so much different. You can call a guy stupid, but that's because he's a guy that's being stupid. But when you call a woman stupid, </s><s>[INST] she's stupid because she's a woman. Yeah, but I honestly think that women are called stupid more than men on the internet. But it's nothing to do, like the attack is not gendered. It's the gender inspires an increased level of attack. [/INST] I feel like it on the internet is gendered. I wish we had data on this. Have you ever heard of the XKCD comics? Yes. It's a really good comic where, and this is something that I've dealt with a lot in my community. There's a guy at a board and he fucks up a math equation and it's like, wow, you suck at math. And then the next panelist is a girl that does it and she fucks it up. And it's like, wow, women suck at math. And there's that feeling that happens. Where when I bring on, I won't use names, but they're like YouTube people that I brought on that have crazy opinions. And when they're men, that person is crazy. Oh my God, he said the crazy stuff. He's so dumb. He's so crazy, he's so stupid. But when it's a woman, it's like, oh my God, why do you always bring dumb women here? Why do so many women on the internet have crazy opinions? There's a different like minority character that has to like stand it and like represent like their whole group </s><s>[INST] where like white men don't typically have to. Speaking of groups versus individuals, yes. But then what I feel happens is then another person from that group comes, another woman comes and people before she says anything [/INST] will already feel like they're ready with that attack. Like for sure. But they're ready for the attack because she's a woman. They're gonna call her she's stupid because she's a woman. Not because she says something but just because she's a woman. </s><s>[INST] So like the group in their brain accumulates all the negative characteristics of the individuals they've met. Not the positive, the negative. And it becomes like this ball of stickiness, and then that becomes the bias for their judgment of a new person that comes. With white men, there's more of a blank slate in terms of bias of how they analyze the person. With any other minority group, they basically make a judgment based on the negative characteristics of the individuals they've met in the past. That leads to a system where you're just harsher [/INST] towards minority groups and towards women. How do you solve that? The most important thing for any problem ever is step one is to be aware of it. If you're not aware of it, then you're hopelessly lost at sea. But yeah, the first thing I like to say, she's like, be aware of it. Like I've had, there's a girl that I've had on recently and she says a lot of, in my opinion, kind of crazy things, but people will use her as like, this is why women shouldn't be here. This is like, she's crazy and she's a woman and blah, blah, blah, but I can put on a guy who's The similarly dumb things, and he's evaluated on his own merits because it's a guy. There's never ever, ever been a case where I brought a stupid guy on stream and everybody's not like, this guy makes me hate men. This guy makes me hate white people. That has never happened, But then there's like other women that come on and it's like, now I know white need cells exist, or I totally understand where red pill ideology comes from. Even if the statements are kind of true, when you're making these observations over and over and over and over again, It damages your ability to individually perceive somebody. And then two people that make the same statements, one can be perceived more harshly just because of that like group bias you've got built up. </s> |