YouSaid · the spoken record

John Carmack

lines on the record: 363
first: 2022-08-04
most recent: 2022-08-04
sittings or episodes: 1
sources: podcast

Every line below is reproduced as it was said and linked to the record it came from. Nothing here is summarised or generated. Directory · Search · Corrections

“Vision matters some people, but it doesn't seem to be a necessary motivator for me. And I think that the process of getting there is usually done. Again, it's like the magic of gradient descent. People just don't believe that just looking locally gets you to all of these spectacular things. That's been, you know, the decades of looking at really some of the smartest people in the world that would just push back forever against this idea that it's not this grand, sophisticated vision of everything, but little tiny steps, local information winds up leading to all the best answers.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“It's been my experience that people that focus on, that don't focus on the here and now right in front of them tend to be less effective. I mean, it's not 100%.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“We are biological creatures that happen stance of evolution. We have innate drives that evolution crafted for survival and passing on of genetic codes. I don't find a lot of value in trying to go much deeper than that. I have my motivations, some of which are probably genetically coded and many of which are contingent on my upbringing and the path that I've had through my life. I don't run into spates of depression or envy or anything that winds up being a challenge and forcing a degree of soul searching with things like that. I seem to be okay, you know, kind of without that”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So, very similar with that last question. I know a lot of people fret about this question a lot, and I just really don't.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And I was thinking to myself at some level that, okay, I mean, I may have a couple more swings at bat with me at full capability, but yes, my mental abilities will decay with age, mostly inevitably. I don't think it's a 0% chance that we will address some of that before it becomes a problem for me. I think exciting medical stuff in the next couple decades. But I do have this kind of vague plan that when I'm not at the top of my game and I don't feel that I'm in a position to put a dent in the world some way, that I'll probably wind up doing some kind of recreational retro programming or I'll work on something that I would not devote my life to now, but I can while away my time as the old man gardening in the code worlds.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“When I decided to work on when I was like nuclear fission or AGI, these are big ticket things that are impact large fractions of the world.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I'm very much focused on what I'm working on right now. I do take that back. There's one aspect where the kind of finiteness of the life does impact me. And that is about thinking about the scope of the problems that I'm working on.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So, I mean, I think I'm an outlier in that where it's, yeah, it doesn't wind up being a real part of my thinking and motivation about things”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Really don't. And you probably always have to take with a grain of salt anything somebody says about fundamental things like that. But I don't think about really aging, impending death, legacy with my children, things like that. And clearly, it seems most of the world does a lot, a lot more than I do.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Your curiosity to find, to kind of like make things useful and valuable to you, even if they don't immediately appear that way.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, and I don't want to tell people that look, this is a good career move just grit your teeth and, you know, and bear it. You want people, you want to, and I do think it is possible sometimes to find the joy in something. Like it might not immediately appeal to you, but I had told people early on in software times that a lot of game developers are in it just because they are so passionate about games, but I was always really more flexible in what appealed to me where I said, I think I could be quite engaged doing operating system work or even database work. I would find the interest in that because I think most things that are significant in the world have a lot of layers and complexity to them and a lot of opportunities hidden within them. So that would probably be the most important thing to encourage to people is that you can, it's like weaponized curiosity. You can deploy”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“People these days have the same relationship with books that I do where they were such a cornerstone for me in so many ways. But that sense that, yeah, I always wanted to know it all. I know I can't. And that was like one of the last things I said. You can't know everything, but you should convince yourself that you can know anything. Any one particular thing, it was created and discovered by humans. You can learn it. You can find out what you need on there.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Trees and the different things that I could do with those systems and just the huge larval phases going through aerospace, just reading bookshelves full of books. I mean, again, that point where I have enough money I can buy all the books I want. It was so valuable where I was terrible with my money when I was a kid. My mom thought I would always be broke because I'd buy my comic books and just be out of money. But it was like all the pizza I want, all the diet coke I want, video games, and then books. And it didn't take that much. As soon as I was making 27K a year, I felt rich and I was just getting all the things that I wanted. But that sense of, you know, books have always been magical to me. And that was one of the things that really made me smile is Andre had said he found when he came over to my house, he said he found my library inspiring just like that. And it was great to see, you know, I still look at him. He's kind of a younger guy. I sometimes wonder if younger.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So, in the earliest days of personal computers, I remember the reference manuals and the very early ones even had schematics of computers in the background, in the back of the books, as well as firmware listings and things. And I could look at that. And at that time, when I was a younger teenager, I didn't understand a lot of that stuff, how the different things worked. I was pulling out the information that I could get, but I always wanted to know all of that. There was like kind of magical information sitting down there. It's like the elder lore that some graybeard wizard is the keeper of. And so I always felt that pull for wanting to know more, wanting to explore the mysterious areas there. And, you know, and that followed right in through all the things that got the value, exploring the video cards leading to the scrolling advantages, you know, exploring some of the academic papers and things, learning about BS.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Where you can't just say, This is my goal in life and just push at that. I mean, you might be able to do that, but you're going to make more total progress if you say, I'm preparing myself with this broad set of tools, and then I'm being aware of all the way things are changing as I move through the world and as the whole world changes around me. And then looking for opportunities to deploy the tools that you've built. And there's going to be more and more of those types of things there where an awareness of what's happening, where the inefficiencies are, what things can be done, what's possible versus what's current practice. And then finding those areas where you can go and make an adjustment and make something that may affect millions or billions of people in the world, make it better.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And that the whole world doesn't need that. You know, most programmers don't need or engineers of any kind don't necessarily need to do that. They need to do the little job that's been parceled out to them, be reliable, let people depend on you, do quality work with all of that. But people that do have an inclination for wanting to know things deeper and learn things deeper. There are just layers and layers of things out there and it's amazing. If you're the right person that is excited about that, the world's never been like this before. It's better than ever. I mean, everything that was wonderful for me is still there and there's whole new worlds to explore on the different things that you can do and that I am, you know, it's hard work. Embrace the grind with it and understand as much as you can and then be prepared for opportunities to present themselves.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So it was kind of fun. I got invited to give the commencement speech back at the, I went to a college for two semesters and dropped out and went on to do my tech stuff, but they still wanted me to come back and give a commencement speech. And I've got that pinned on my Twitter account. I still feel good about everything that I said there. And my biggest point was that the path for me might not be the path for everyone. And in fact, the advice, the path that I took and even the advice that I would give based on my experience and learnings probably isn't the best advice for everyone because what I did was all about this knowledge in depth. It was about not just having this surface level ability to make things do what I want, but to really understand them through and through, to let me do the systems engineering work and sometimes find these inefficiencies that can be bypassed.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, because nobody's any martial art is coming out thinking, I'm the best in the world at anything because everybody loses.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And in the most extreme case, when you tap to a choke, you are basically living because somebody lets you live. And that is one of those, if you think about it, that is a closer brush with mortality than most people consider.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“It was interesting where I did go to the Codicon in Japan, kind of the birthplace of judo and everything. And I remember I rolled with one old guy. I didn't start standing, just started on groundwork Niwaza. And it was striking how different it was from Carlos. He was still, he was better than me, and he got my arm. And I had to tap there, but it was a completely different style where I just felt like I could do nothing. He was just...”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And of course, when I worked with good guys like Carlos, it's like just the whole flow like water thing is real. And he's just like.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I'd be dangerous to anyone for the first five minutes, but then sometimes after that I'm already dead. And I knew it was terrible for me because it made the, it meant I got less training time with all of that when you go and you just gas out relatively quickly there. And I like to think that I would be better about that where after I gave up Judo, I started doing the half marathons and tough butters and things like that. And so when I did go back to the local Judo Kai club, I thought it's like, oh, I should have better cardio for this because I'm a runner now and I do all of this and didn't work out that way. It was the same old thing where just push really hard, strain really hard.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I still pushed kind of harder than I should. I mean, that was, I was one of those people that I'm on the smaller side for a lot of the people competing. And I would go with all the big guys and I'd go hard and I'd push myself a lot. And that would be one of those where.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“When I went back a few years ago, it was good to see that I whipped myself back into reasonable shape about doing the basic grappling, but I know there was a ton of the subtleties that were just, that were gone, but could probably be brought back reasonably quickly.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And I do think about that more as kind of an older person considering the martial arts where I can remember the very earliest days getting back into judo and I'm like teach me submissions right now. You know, it's like learn the arm bar, learn the choke. But as you get older, you start thinking more about it's like, okay, I really do want to learn the entire canon of judo. It's like, are all the different things there and like all the different approaches for it, not just the, you know, if you want to compete, there's just a handful of things you learn really, really well. But sometimes there's interest in learning a little bit more of the scope there and figuring some things out from at one point I had wasn't exactly a spreadsheet, but I did have a big long text file with like, here's the things that I learned and here are like ways you chain this together.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“But if you're going to a judo club or something, you've got to fix it. It's going to be seven o'clock or whatever, 10 o'clock on Saturday. Although I talked about this a little bit when I was on Rogan. And shortly after that, Carlos Machado did reach out and I had trained with him for years back in the day. And he was like, hey, we've got kind of a small private club with a bunch of executive type people.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah. And I was thinking, yeah, I guess we're in the executive division if you're over 50 or over 45 or something. And it's not out of the question that I go back at some point and do some of this. But again, I'm just reorganizing my life around more focus. Probably not going to happen. I'm pushing my exercise around to give me longer uninterrupted intellectual focus time, pushing it to the beginning or the end of the running and stuff.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I've gone a couple times in the last decade trying to get my kids into it a little bit. I didn't really stick with any of them, but it was fun to get back on the mats. It really hurts for a while when you haven't gone for a while, but I still debate this pretty constantly. My brother's only a year younger than me, and he's going kind of hard in jiu-jitsu right now. And he was just, he won a few medals at the last tournament he was at P.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“You know, the sense that I, you know, working out with the group and having the guys that you're beating each other up with it, but you just feel good coming out of it. And I can remember those driving home aching in various ways and just thinking it's like, oh, that was really great. And it's mixing with a bunch of people that had nothing to do with any of the things that I worked with. Every once in a while, someone would be like, oh, you're the Doom guy. But for the most part, it was just different slice of life. A good thing. And I made the call when I was 40. That's like maybe I'm getting a little old for this. I had separated a rib and tweaked a few things. And I got out without any really bad injuries. And was like, have I dodged enough bullets? Should I, you know, should I hang it up? I went back.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“But there's also that sense that I was not a sports guy. I did do wrestling in junior high, and I often wish that I think I would have been good for me if I'd carried that on into high school and had a little bit more of that. I mean, it's like I filch a little bit of wrestling vibe with what was going on about embracing the grind and like that push that I associate with the wrestling team that I in hindsight I wish I had gone through that and pushed myself that way. But even getting back into judo and jiu-jitsu in my mid-30s as usually the old man on the mat with that, there was still the...”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I mean, I was a competent club player in judo and grappling. I mean, I was by no means any kind of a superstar, but it was, I went through a few phases with it where I did some when I was quite young, a little bit more when I was 17, Ion. And then I got into it kind of seriously in my mid-30s. And I went pretty far with it. And I was pretty good at some of the things that I was doing. And I did appreciate it quite a bit where on the one hand, it's always if you're going to do exercise or something, it's a more motivating form of exercise. Someone is crushing you, you are motivated to do something about that, to up your attributes and be better about getting out of that attributes.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, like I say, I sketch with structs and things when I'm just thinking about a problem. I'm thinking in some degree of code.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“He was very much a programmer's programmer that was doing machine learning work rather than it's a different feel than an academic where you can see it in paper sometimes where somebody that's really a mathematician or a statistician at heart and they're they're doing something with machine learning but Andre is about getting something done and you could see it in like all of his earliest approaches to it's like okay here's how reinforcement learning works here's how recurrent neural networks work here's how transformers work here's how crypto works and i am and yeah it's just he's a hacker's you know one of his old posts was like a hacker's guide to machine learning yeah and you know he deprecated that and said don't really pay attention to what's in here but it's that thought that uh that carries through in a lot of it where it is that back again to that hacker mentality and the hacker”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So, yeah, first of all, Andre is awesome. I learned a lot when I was going through my larval phase from his blog posts and his Stanford course and super valuable. I got to meet him first a couple years ago when I was first kind of starting off on my gentleman scientist bit. And just a couple months ago, when he went out on his sabbatical, he stopped by in Dallas and we talked for a while and I had a great time with him. And then when I heard he actually left Tesla, I did, of course, along with a hundred other people say, hey, if you ever want to work with me, it would be an honor. So he thinks that he's going to be doing this educational work, but I think someone's going to make him an offer he can't refuse before he gets too far along on it.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“This investment. Investors that are going to expect something of me. Now, we've all had the conversation that this is a low probability long-term bet. It's not something that there's a million things I could do that I would have line of sight on the value proposition for. This isn't that. I think there are unknown unknowns in the way, but it's one of these things that it's hyperbol, but it's potentially one of the most important things humans ever do. And it's something that I think is within our lifetimes, if not within a decade, to happen. So, yeah, this is just now happening, like term sheet, like the ink is barely, HertLink's barely dry on.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“But so the next big thing, though, is I decided that I was going to take some investor money because I have an overactive sense of responsibility about other people's money. And it's like I don't want, I mean, a lot of my push and my passionate entreaties for things at Meta are it's like, I don't want Sukt to have wasted his money investing in Oculus. I want it to work out. I want it to change the world. I want it to be worth all of this time, money, and effort going into it. And I expect that it's going to be like that with my company where...”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“that I've done that are kind of tactics for myself to make me more effective. Like one thing I noticed I was not doing well is I had a Google Cloud account with to get GPUs there and 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. I mean, 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 like 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. You know, 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 experiment with different”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“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 the 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 a million dollars 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”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“It 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 Suscover 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 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 eventually 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.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Now, it was a little bit, 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'd gone and written it all in C just to make sure I understood back propagation at the lowest level. And my nuts and bolts approach. But after Sam approached me,”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“And 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 find something to talk about.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“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 that 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 Einstein's and Bohr's 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, you know, 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 ago for your AI.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Nice position there. I only have to work one day a week for my consulting role, but I was engaging every day. I'd still be like my computer's there. I'd be going and checking the workplace and notes and testing different things and communicating with people. But I did make the decision recently that I'm going to get serious. I'm still going to keep my ties with Meta, but I am seriously going for the AGI side of things.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“The contingency of your path to knowledge and talking about the associations and the context that you have with them, where people that learn in the same path will have similar thought processes. And I think it's useful that I come at this from a different background, different history than the people that have had the largely academic backgrounds for this, where I have huge blind spots that they could easily point out, but I have a different set of experiences in history and approaches to problems and systems engineering that might turn out to be useful and I can afford to take that bet where I'm not going to be destitute. I have enough money to fund myself working on this for the rest of my life. But what I was finding is that I was still not committing where I had a foot firmly in the VR and meta side of things where in theory I've got a very”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“I mean, maybe people are still think that it's an unapproachable problem, that it's kind of beyond their ability to get a wrench on and have some effect on like whatever startups they've run before. But that was my kind of like with all the stuff I've learned, whether it's gaming, aerospace, whatever, I go through a larval phase where I'm like, okay, I'm sucking up all of this information, trying to see, is this something that I can actually do? Is this something that's practical to devote a large chunk of my life to? And I've gone through that with the AI machine learning space of things. And I think I've got my arms around it. I've got the measure of it where some of the most brilliant people in the world are working on this problem, but nobody knows exactly the path that it's going on. We're throwing a lot of things at the wall and seeing what sticks. But I have another interesting thing, just learning about all of this.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So I literally just this week signed a term sheet to take some investment money for my company where the last two years I had backed off from Meta and I was still doing my consulting CTO role there, but I had styled it as I was going to take the Victorian gentleman scientist route where I was going to be the wealthy person that was going to go pursue science and learn about this and do experiments. And honestly, I'm surprised there aren't more people like that that are like me, technical people that made a bunch of money and are interested in some of these possibly the biggest leverage point in human history. I mean, I know of, I've heard of a couple organizations that are basically led by one rich techie guy that gets a few people around him to try to work on this. But I'm surprised that there's not more, that there aren't like a dozen of them.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Yeah, I mean, it's back to the very earliest days of supercomputers. You still have the balance between bandwidth storage and computation. And sometimes they're easier to get one or the other, but it's been remarkably constant across all those years that you still need all three.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“Real fundamental issues there when you start getting down to taking an actual problem and putting it on an abstract machine like that That has not worked out well in practice. And the idea that there was always been easy to come up with ways to get compute faster, say more flops or more giga-ops or whatever there, that's usually the easy part, but you then have interconnect and then memory for what goes into it. And when you talk about saying, well, cell phones, well, you're limited to like a 5G connection or something on that. And if you say how if you take your your calculation and you factor it across a million cell phones instead of a thousand GPUs in a warehouse, you might be able to have some kind of a substrate like that, but it could be operating then at 1 1000th the speed. And so yes, you could get you could have an AGI working there, but it wouldn't be a real-time AGI. It would be something that is operating at really a snail's pace, you know, much, much slower than kind of human level thought for things.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“You are, even if you have perfect hacking ability that take over the world in an instant sort of thing, just isn't plausible at all. And even if you had access to all of the resources, these are going to be specialized systems where you're going to wind up with something that is architected around exactly this chip with this interconnect. And it's not just going to be able to be plopped somewhere else. Now, interestingly, it is going to be something that the entire code for all of it will easily fit on a thumb drive. That's total spy movie thriller sorts of things where you could have, hey, we crack the secret to AGI and it fits on this thumb drive and anyone could steal it. Now they're still going to have to build the right data center to deploy it and have the right kind of life experience curriculum.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source
“So, it is my belief that we're going to start off with something that requires thousands of GPUs. And I don't know if you've tried to go get a thousand GPU instance on a cloud anytime recently, but these are not things that you can just go spin up hundreds of. There are real challenges to, I mean, these things are going to take data centers and data centers take years to build. The last few years, we've seen a few of them kind of coming up, going in different places. They're big engineering efforts. You can hear people bemoan about the fact that I, you know, oh, the network was wired all wrong and it took them a month to go unwire it and rewire it the right way. These aren't things that you can just magic into existence. And the ideas of like the old tropes about it's going to escape onto the internet and take over other systems. There's the fast takeoff ones are clearly nonsense because you just can't open TCP connections above a certain rate no matter how smart.”
2022-08-04 · Lex Fridman Podcast · #309 – John Carmack: Doom, Quake, VR, AGI, Programming, Video Games, and Rockets · IDENTIFIED FROM THE TRANSCRIPT · source