Skip to content
Lex Fridman PodcastLex Fridman Podcast

George Hotz: Tiny Corp, Twitter, AI Safety, Self-Driving, GPT, AGI & God | Lex Fridman Podcast #387

George Hotz is a programmer, hacker, and the founder of comma-ai and tiny corp. Please support this podcast by checking out our sponsors: - Numerai: https://numer.ai/lex - Babbel: https://babbel.com/lexpod and use code Lexpod to get 55% off - NetSuite: http://netsuite.com/lex to get free product tour - InsideTracker: https://insidetracker.com/lex to get 20% off - AG1: https://drinkag1.com/lex to get 1 month supply of fish oil TRANSCRIPT: https://lexfridman.com/george-hotz-3-transcript EPISODE LINKS: George's Twitter: https://twitter.com/realgeorgehotz George's Twitch: https://twitch.tv/georgehotz George's Instagram: https://instagram.com/georgehotz Tiny Corp's Twitter: https://twitter.com/__tinygrad__ Tiny Corp's Website: https://tinygrad.org/ Comma-ai's Twitter: https://twitter.com/comma_ai Comma-ai's Website: https://comma.ai/ Comma-ai's YouTube (unofficial): https://youtube.com/georgehotzarchive Mentioned: Learning a Driving Simulator (paper): https://bit.ly/42T6lAN PODCAST INFO: Podcast website: https://lexfridman.com/podcast Apple Podcasts: https://apple.co/2lwqZIr Spotify: https://spoti.fi/2nEwCF8 RSS: https://lexfridman.com/feed/podcast/ Full episodes playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4 Clips playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOeciFP3CBCIEElOJeitOr41 OUTLINE: 0:00 - Introduction 1:39 - Time is an illusion 11:18 - Memes 13:55 - Eliezer Yudkowsky 26:19 - Virtual reality 32:38 - AI friends 40:03 - tiny corp 53:24 - NVIDIA vs AMD 56:21 - tinybox 1:08:30 - Self-driving 1:23:09 - Programming 1:31:06 - AI safety 1:56:03 - Working at Twitter 2:33:46 - Prompt engineering 2:39:42 - Video games 2:55:57 - Andrej Karpathy 3:06:02 - Meaning of life SOCIAL: - Twitter: https://twitter.com/lexfridman - LinkedIn: https://www.linkedin.com/in/lexfridman - Facebook: https://www.facebook.com/lexfridman - Instagram: https://www.instagram.com/lexfridman - Medium: https://medium.com/@lexfridman - Reddit: https://reddit.com/r/lexfridman - Support on Patreon: https://www.patreon.com/lexfridman

Lex FridmanhostGeorge Hotzguest
Jun 29, 20233h 8mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:001:39

    Introduction

    1. LF

      What possible ideas do you have for the w- how human species ends?

    2. GH

      Sure. So I think the most obvious way to me is wire heading. We end up amusing ourselves to death. We end up all staring at that infinite TikTok and forgetting to eat. Maybe, maybe it's even more benign than this. Maybe we all just stop reproducing. Now, to be fair, it's probably hard to get all of humanity.

    3. LF

      Yeah. The interesting thing about humanity is the diversity in it.

    4. GH

      Oh, yeah.

    5. LF

      Organisms in general. There's a lot of weirdos out there.

    6. GH

      Well-

    7. LF

      Two of them are sitting here.

    8. GH

      I mean, diversity in humanity is-

    9. LF

      With due respect. (laughs)

    10. GH

      (laughs) I wish I was more weird.

    11. LF

      The following is a conversation with George Hotz, his third time on this podcast. He's the founder of Comma.ai that seeks to solve autonomous driving, and is the founder of a new company called TinyCorp that created TinyGrad, a neural network framework that is extremely simple, with the goal of making it run on any device by any human easily and efficiently. As you know, George also did a large number of fun and amazing things, from hacking the iPhone to recently joining Twitter for a bit as an "intern" in quotes, making the case for refactoring the Twitter code base. In general, he's a fascinating engineer and human being, and one of my favorite people to talk to. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description. And now, dear friends, here's George Hotz.

  2. 1:3911:18

    Time is an illusion

    1. LF

      You mentioned something in a stream about the philosophical nature of time. So, uh, let's start with a wild question. Do you think time is an illusion?

    2. GH

      You know-

    3. LF

      (laughs)

    4. GH

      ... I sell phone calls, uh, to Comma for $1,000. Uh, and some guy called me, and, uh, like, you know, it's $1,000, so you can talk to me for, for half an hour. And he's like, "Uh, yeah, okay. So like, time doesn't exist, and I really wanted to share this with you." I'm like, "Well, what do you mean time doesn't exist," right? Like, I think time is a useful model whether it exists or not, right? Like, does quantum physics exist? Well, it doesn't matter. I- i- it's about whether it's a useful model to describe reality. Is time maybe compressive?

    5. LF

      Do you think there is an objective reality or is everything just useful models? Like underneath it all, is there an actual thing that we're constructing models for?

    6. GH

      I don't know.

    7. LF

      I was hoping you would know.

    8. GH

      I don't think it matters.

    9. LF

      I mean, this kind of connects to the models we construct of reality with machine learning, right?

    10. GH

      Sure.

    11. LF

      Like, is it just nice to have useful approximations of the world such that we can do something with it?

    12. GH

      So there are things that are real. Kolmogorov complexity is real.

    13. LF

      Yeah.

    14. GH

      Yeah. The compressive-

    15. LF

      Math.

    16. GH

      Math is real, yeah.

    17. LF

      (laughs) Should be a T-shirt.

    18. GH

      And I think hard things are actually hard. I don't think P equals NP.

    19. LF

      Ooh, strong words.

    20. GH

      Well, I think that's the majority. I do think factoring is in P, but...

    21. LF

      I don't think you're the person that follows the majority in all walks of life, so let's go-

    22. GH

      For that, for that one, I do.

    23. LF

      Yeah. In theoretical computer science, you, you're, you're one of the sheep. (laughs)

    24. GH

      (laughs)

    25. LF

      All right. But to you, uh, time is a useful model.

    26. GH

      Sure.

    27. LF

      Hmm. What were you talking about on the stream with time? Are you made of time?

    28. GH

      If I remembered half the things I said on stream.

    29. LF

      Ah.

    30. GH

      Someday, someone's gonna make a model of all of it, and it's gonna come back to haunt me.

  3. 11:1813:55

    Memes

    1. LF

      Uh, do you think we have ideas, or ideas have us?

    2. GH

      I think that we're gonna get super scary memes once the AIs actually are superhuman.

    3. LF

      Ooh.

    4. GH

      Like, the-

    5. LF

      You think AI will generate memes?

    6. GH

      Of course.

    7. LF

      You think it'll make humans laugh?

    8. GH

      I think it's worse than that. So, um, Infinite Jest, uh, it's introduced in the first 50 pages, is about a tape that you, uh, once you watch it once, you only ever want to watch that tape. Um, in fact, you want to watch the tape so much that someone says, "Okay, here's a hacksaw. Cut off your pinky and then I'll let you watch the tape again," and you'll do it. Uh, so we're actually gonna build that, I think. But it's not gonna be one static tape. I think the human brain is too complex to be stuck in one static tape like that. If you look at, like, ant brains, maybe they can be stuck on a static tape. But we're going to build that using generative models. We're going to build the TikTok that you actually can't look away from.

    9. LF

      So TikTok is already pretty close there with the generation that's done by humans.

    10. GH

      Yeah.

    11. LF

      The algorithm is just doing their recommendation, but if it's do- if the algorithm is also able to do the generation...

    12. GH

      Well, it's a question about how much intelligence is behind it, right?

    13. LF

      Mm-hmm.

    14. GH

      So the content is being generated by, let's say, one humanity worth of intelligence. And you can quantify a humanity, right? That's a ... you know, it's- it's exaflops, yottaflops, uh, but you can quantify it. Once that generation is being done by a hundred humanities, you're done.

    15. LF

      (sighs) So it's actually scale that's the problem, but also speed. Yeah. And what if it's sort of manipulating the very limited human dopamine engine for porn? Imagine it's just TikTok but for porn.

    16. GH

      Yeah.

    17. LF

      That's like a brave new world.

    18. GH

      I don't even know what it'll look like, right? Like again, y- you can't imagine the behaviors of something smarter than you. But a super intelligent and- and- and- and- and agent that just dominates your intelligence so much will be able to completely manipulate you.

    19. LF

      Is it possible th- that it won't really manipulate, it'll just move past us? It'll just kinda, uh, exist the way water exists or the air exists?

    20. GH

      You see? And that's the whole AI safety thing.

    21. LF

      (laughs)

    22. GH

      It's not the machine that's gonna do that. It's other humans using the machine that are gonna do that to you.

    23. LF

      Yeah. 'Cause the machine is not interested in hurting humans. It's just-

    24. GH

      The machine, the machine is a machine.

    25. LF

      Yeah.

    26. GH

      But the human gets the machine, and there's a lot of humans out there very interested in manipulating you.

  4. 13:5526:19

    Eliezer Yudkowsky

    1. GH

    2. LF

      Well, let me bring up Eliezer Yudkowsky, who recently sat where you're sitting. (laughs) He thinks that AI will almost assuredly kill everyone. Do you agree with him or not?

    3. GH

      Yes, but maybe for a different reason.

    4. LF

      (laughs) Okay. W- (laughs) and then I'll try to, uh, get you to find hope where we could find a no to that answer. But why yes?

    5. GH

      Okay. Why didn't nuclear weapons kill everyone?

    6. LF

      That's a good question.

    7. GH

      I think there's an answer. I think it's actually very hard to deploy nuclear weapons tactically.

    8. LF

      Mm-hmm.

    9. GH

      It's very hard to accomplish tactical objectives. Great, I can nuke their country and have a irradiated pile of rubble. I don't want that.

    10. LF

      Why not?

    11. GH

      Why don't I want an irradiated pile of rubble?

    12. LF

      Yeah.

    13. GH

      For all the reasons no one wants an irradiated pile of rubble. (laughs)

    14. LF

      Oh, 'cause you can't use that land for, uh, for resources. You can't populate the land.

    15. GH

      Yeah. W- what you want a- a- a total victory in a war is not usually the ir- irradiation and eradication of the people there. It's the subjugation and domination of the people.

    16. LF

      Mm-hmm. Okay. So you can't use this strategically, tactically in a war-

    17. GH

      Yeah.

    18. LF

      ... to help you, to- so to help, uh, gain a military advantage. It's all complete destruction. All right.

    19. GH

      Yeah.

    20. LF

      But there's egos involved. It's still surprising. Still surprising that nobody pressed the big red button.

    21. GH

      It's somewhat surprising. But you see, it's the little red button that's gonna be pressed with AI that's gonna ... you know, and that's why we die. It's- it's not because the AI, if there's anything in the nature of AI. It's just the nature of humanity.

    22. LF

      What's the algorithm behind the little red button? Or like, what- what- what possible ideas do you have for the w- how human species ends?

    23. GH

      Sure. So I think the most, uh, obvious way to me is wire heading. We end up amusing ourselves to death. We end up all staring at that infinite TikTok and forgetting to eat. Maybe- maybe it's even more benign than this. Maybe we all just stop reproducing. You know, to be fair, it's probably hard- hard to get all of humanity.

    24. LF

      Yeah.

    25. GH

      Yeah. (laughs) It probably-

    26. LF

      This is always go ... Like the- the interesting thing about humanity is the diversity in it.

    27. GH

      Oh, yeah.

    28. LF

      O- organisms in general. There's a lot of weirdos out there.

    29. GH

      Well-

    30. LF

      Two of them are sitting here.

  5. 26:1932:38

    Virtual reality

    1. GH

      time.

    2. LF

      You like virtual reality?

    3. GH

      I love it.

    4. LF

      Do you want to live there?

    5. GH

      Yeah.

    6. LF

      Yeah. Part of me does too. How far away are we, do you think?

    7. GH

      Judging from what you can buy today? Far. Very far.

    8. LF

      I gotta tell you that I had the experience of, uh, Meta's Codec avatar, where it's a ultra-high resolution scan. It looked real.

    9. GH

      I mean, the headsets just are not quite at, like, eye resolution yet. I, I haven't put on any headset where I'm like, "Oh, I, this could be the real world." Whereas, when I put good headphones on, audio is there. And like, we, we can reproduce audio that I'm like, "I'm actually in a jungle right now." I, I, if I close my eyes, I can't tell I'm not.

    10. LF

      Yeah. But then, there's also smell and all that kind of stuff.

    11. GH

      Sure.

    12. LF

      I don't know. I... The, the power of imagination or the power of the, the mechanism in the human mind that fills the gaps, that kind of reaches and wants to make the thing you see in the virtual world real to you, I believe in that power.

    13. GH

      Or humans want to believe.

    14. LF

      Yeah. Like, what, what if you're lonely? What if you're sad? What if you're really struggling in life and here's a world where you don't have to struggle anymore?

    15. GH

      Humans want to believe so much that people think the large language models are conscious. That's how much humans want to believe.

    16. LF

      Strong words. He's throwing left and right hooks.

    17. GH

      (laughs)

    18. LF

      Uh, why do you think large language models are not conscious? Do you think-

    19. GH

      I don't think, I don't think I'm conscious.

    20. LF

      Oh. So what is consciousness then, George Hotz?

    21. GH

      It's like, what it seems to mean to people, it's just like a word that atheists use for souls.

    22. LF

      Sure. But that doesn't mean that soul is not an interesting word.

    23. GH

      If consciousness is a spectrum, I'm definitely way more conscious than the large language models are. I think the large language models are less conscious than a chicken.

    24. LF

      When is the last time you seen a chicken?

    25. GH

      Uh, in Miami, like, a couple months ago.

    26. LF

      How... Uh, no. Like a living chicken.

    27. GH

      There's living chickens walking around Miami. It's crazy.

    28. LF

      Like, on the street?

    29. GH

      Yeah.

    30. LF

      Like a chicken?

  6. 32:3840:03

    AI friends

    1. GH

      lot of value in it. Look, I just started my second company. My third company will be AI girlfriends.

    2. LF

      (laughs)

    3. GH

      No, like, I mean it.

    4. LF

      I want to find out what your fourth company is after that-

    5. GH

      Oh, wow. (laughs)

    6. LF

      ... because I think o- once you have AI girlfriends, it's, uh... Oh boy, does it get interesting. Well, maybe let's go there. I mean, the relationships with AI, that's creating human-like organisms, right? And part of being human is being conscious, is being, uh, having the capacity to suffer, having the capacity to experience this life richly in such a way that you can empathize, the AI s- can empathize with you and you can empathize with it, or you can project your, uh, anthropomorphic sense of what the other entity is experiencing. And, and an AI model would need to, um, yeah, to create that experience inside your mind. And it doesn't seem that difficult.

    7. GH

      Y- yeah, but... Okay. So here's where it actually gets totally different, right? When you interact with another human, you can make some assumptions.

    8. LF

      Yeah.

    9. GH

      When you interact with these models, you can't. You can make some assumptions that that other human experiences suffering and pleasure in a pretty similar way to you do. The golden rule applies.

    10. LF

      Mm-hmm.

    11. GH

      With an AI model, this isn't really true, right? Th- these, these large language models are good at fooling people because they were trained on a h- uh, a whole bunch of human data and told to mimic it.

    12. LF

      Yep. But if, if the AI system says, "Hi, my name is Samantha," uh, it has a backstory.

    13. GH

      Yeah.

    14. LF

      Went to college here and there.

    15. GH

      Yeah.

    16. LF

      Maybe it'll integrate that's an AI system.

    17. GH

      I made some chatbots. I gave them backstories. It was lots of fun. I was so happy when LLaMA came out.

    18. LF

      Yeah. Uh, we'll talk about LLaMA. We'll talk about all that, but like, you know, the rock with a smiley face.

    19. GH

      Yeah.

    20. LF

      Why... That's pr- it seems pretty natural for, for you to anthropomorphize that thing and then start dating it, and before you know it, you're married and have kids.

    21. GH

      With a rock? (laughs)

    22. LF

      (laughs) With a rock.

    23. GH

      Oh.

    24. LF

      There's pictures on Instagram with you and a rock and a smiley face.

    25. GH

      To be fair, like, you know, something that people generally look for when they're looking for someone to date is intelligence in some form, and the rock doesn't really have intelligence. Only a pretty desperate person would date a rock.

    26. LF

      I think we're all desperate deep down.

    27. GH

      Oh, not rock level desperate.

    28. LF

      (laughs) All right. Uh, not rock level desperate, but AI level desperate. I don't know. I think all of us have a deep loneliness. It just feels like the language models are there.

    29. GH

      Oh, I agree. And you know what? I won't even say this so cynically. I will actually say this in a way that, like, I want AI friends. I do.

    30. LF

      Yeah.

  7. 40:0353:24

    tiny corp

    1. LF

      All right, let's, uh, before we go to company number three and company number four, let's go to company number two.

    2. GH

      All right.

    3. LF

      TinyCorp. Possibly one of the greatest names of all time for a company.

    4. GH

      (laughs)

    5. LF

      Uh, you've launched a new company called TinyCorp that leads the development of TinyGrad. What's the origin story of TinyCorp and TinyGrad?

    6. GH

      I started TinyGrad as a, like a toy project just to teach myself, okay, like, what is a convolution? Uh, what are all these options you can pass to them? What is the derivative of a convolution, right? Very similar to, uh, uh, Karpathy wrote MicroGrad.

    7. LF

      Mm-hmm.

    8. GH

      Um, very similar. And then I started realizing, I started thinking about, like, AI chips.

    9. LF

      Mm-hmm.

    10. GH

      I started thinking about chips that run AI and I- I was like, "Well, okay, this is going to be a- a really big problem. If NVIDIA becomes a monopoly here, um, how long before NVIDIA's nationalized?"

    11. LF

      Hmm. So you, uh... One of the reasons to sta- start TinyCorp is to challenge NVIDIA?

    12. GH

      It's not so much to challenge NVIDIA. I- I actually, I- I like N- NVIDIA and it's to make sure power stays decentralized.

    13. LF

      Yeah. And here, it's, uh, computational power.

    14. GH

      Mm-hmm.

    15. LF

      And to you, NVIDIA is kinda locking down the computational power of the world.

    16. GH

      If NVIDIA becomes just, like, 10X better than everything else, you're giving a big advantage to somebody who can secure NVIDIA as a resource.

    17. LF

      Yeah.

    18. GH

      In fact, if Jensen watches this podcast, he may wanna consider this. He may wanna consider making sure his company's not nationalized.

    19. LF

      D- do you think that's an actual threat?

    20. GH

      Oh, yes.

    21. LF

      No, but there's so much, uh, you know, there's AMD.

    22. GH

      Mm-hmm. So we have NVIDIA and AMD, great.

    23. LF

      All right. But you don't- you don't think there's like a push towards, like, selling? Like, Google selling TPUs or something like this, you don't think there's a push for that?

    24. GH

      Have you seen it? Google loves to rent you TPUs.

    25. LF

      It doesn't... You can't buy it at Best Buy?

    26. GH

      No.

    27. LF

      Hm.

    28. GH

      So I started work on a, uh, on a chip. I was like, "Okay, what's it gonna take to make a chip?" And my first notions were all completely wrong about why, about like how you could improve on GPUs. Uh, and I will take this, this is from, uh, Jim Keller, on your podcast.

    29. LF

      Mm-hmm.

    30. GH

      Uh, and this is one of my absolute favorite descriptions of computation.Um, so there's three kinds of computation paradigms that are common in the world today. Uh, there's CPUs, and CPUs can do everything. CPUs can do add and multiply, they can do load and store, and they can do compare and branch.

  8. 53:2456:21

    NVIDIA vs AMD

    1. LF

      What's the hope that AMD has? I mean, you did a build with AMD recently that I saw. Uh, how does the, uh, uh, the 7900 XTX compare to the RTX 4090 or 4080?

    2. GH

      Oh. Well, let's start with the fact that the 7900 XTX kernel drivers don't work. And if you run demo apps in loops, it panics the kernel.

    3. LF

      Okay. So this is a software issue.

    4. GH

      Lisa Su responded to my email.

    5. LF

      Oh.

    6. GH

      I reached out. I was like, "This is, you know... Really?"

    7. LF

      Yeah.

    8. GH

      Like I understand if your seven by seven transposed Winograd conv is slower than NVIDIA's, but literally when I run demo apps in a loop-

    9. LF

      Yeah.

    10. GH

      ... the kernel panics?

    11. LF

      So just adding that loop?

    12. GH

      Yeah, I just, I just literally took their demo apps and wrote like, "While true; do the app; done" in a bunch of screens.

    13. LF

      Mm-hmm.

    14. GH

      Right? This is like, like the most primitive fuzz testing.

    15. LF

      Why do you think that is? They're just not seeing a market in the, in, um, machine learning?

    16. GH

      They're changing. They're trying to change. They're trying to change. And I had a pretty positive interaction with them this week. Last week I, I went on YouTube, I was just like, "That's it. I give up on AMD." Like this is the... Their driver doesn't even like, I'm not gonna, I'm not gonna... You know, I'm, I'll, I'll go with Intel GPUs, right? Intel GPUs have better drivers.

    17. LF

      (inhales) So you're kind of spearheading the diversification of, uh, GPUs?

    18. GH

      Yeah. And I'd like to extend that diversification to everything. I'd like to diversify the, right, the more... My central thesis about the world is there's things that centralize power and they're bad, and there's things that decentralize power and they're good. Everything I can do to help decentralize power, I'd like to do.

    19. LF

      So you're really worried about the centralization of NVIDIA? That's interesting. And you don't have a fundamental hope for the- the pro- proliferation of ASICs, uh, e- except in the cloud?

    20. GH

      I'd like to help them with software. No, actually there's only, the only ASIC that is remotely successful is Google's TPU.

    21. LF

      Yeah.

    22. GH

      And the only reason that's successful is because Google wrote a machine learning framework. I- I think that you have to write a competitive machine learning framework in order to be able to build an ASIC.

    23. LF

      Hmm. You think Meta with PyTorch builds a competitor?

    24. GH

      I hope so.

    25. LF

      Okay.

    26. GH

      They have one. They have an internal one.

    27. LF

      Internal. I mean, uh, public facing with a nice cloud interface and so on.

    28. GH

      I don't want a cloud.

    29. LF

      You don't like cloud?

    30. GH

      I don't like cloud.

  9. 56:211:08:30

    tinybox

    1. LF

      AMDs. Uh, well, let me ask, sort of on a tangent to ask you, what, what, um, you've built quite a few PCs. What's your advice on how to build a good custom PC for, uh, let's say for the different applications that you use, for gaming, for, uh, machine learning?

    2. GH

      Well, you shouldn't build one. You should buy a box from the Tiny Corp.

    3. LF

      I heard rumors, whispers about this box in the Tiny Corp. What's- what's this thing look like? What is, what is it, what is it called?

    4. GH

      Uh, it's called the Tiny Box.

    5. LF

      Tiny Box? Of course it is.

    6. GH

      Um, it's $15,000.

    7. LF

      Yep.

    8. GH

      And it's almost a petaflop of compute. It's over 100 gigabytes of GPU RAM. It's over five terabytes per second of GPU memory bandwidth. Uh, I'm gonna put, like, four nVMes in- in RAID. You're gonna get, like, 20, 30 gigabytes per second of drive read bandwidth. I'm gonna, I'm gonna build, like, the best deep learning box that I can that plugs into one wall outlet.

    9. LF

      Okay. Can you go through those specs again a little bit from your m- from memory?

    10. GH

      Yeah. So it's almost a petaflop of compute.

    11. LF

      So AMD, Intel?

    12. GH

      Today, I'm leaning toward AMD.

    13. LF

      Mm-hmm.

    14. GH

      Um, but we're pretty agnostic to the type of compute. The- the- the main limiting spec is a 120 volt, 15 amp circuit.

    15. LF

      (laughs) Okay.

    16. GH

      Well, I mean it, because in order to, like, like there's a plug over there.

    17. LF

      Mm-hmm.

    18. GH

      All right? You have to be able to plug it in. Um, we're also gonna sell the tiny rack, which, like, what's the most power you can get into your house without arousing suspicion? Uh, and one of the, one of the answers is an electric car charger.

    19. LF

      Wait, where does that rack go?

    20. GH

      Your garage.

    21. LF

      Interesting. The car charger.

    22. GH

      A wall outlet is about 1,500 watts. A car charger is about 10,000 watts.

    23. LF

      Is it? (laughs) What is the most amount of power you can get your hands on without arousing suspicion?

    24. GH

      That's right.

    25. LF

      George Hotz. Okay. Uh (laughs) so the- the tiny box, and you said nVMes and RAID.

    26. GH

      Yeah.

    27. LF

      Uh, I forget what you said about memory, all that kind of stuff. Okay.

    28. GH

      So-

    29. LF

      Uh, what about what GPUs?

    30. GH

      Again-

  10. 1:08:301:23:09

    Self-driving

    1. GH

    2. LF

      Since you mentioned OpenPilot, I'd love to get an update in the, uh, company number one, Comma.ai world. How are things going there in the development of, uh, semi-autonomous driving?

    3. GH

      You know, almost no one talks about FSD anymore, and even less people talk about OpenPilot.

    4. LF

      Mm-hmm.

    5. GH

      We've solved the problem. Like, we solved it years ago.

    6. LF

      What's the problem exactly?

    7. GH

      Well, how do you-

    8. LF

      Or what, what, what does solving it mean?

    9. GH

      Solving means how do you build a model that, uh, outputs a human policy for driving?

    10. LF

      Mm-hmm.

    11. GH

      How do you build a model that given our, you know, reasonable set of sensors outputs a human policy for driving? Uh, so you have, you know, companies like Waymo and Cruise which are hand-coding these things that are like quasi-human policies.

    12. LF

      Hm.

    13. GH

      Then you have Tesla and maybe even to more of an extent, Comma, asking, "Okay, how do we just learn the human policy from data?" The big thing that we're doing now, and we just put it out on Twitter. At the beginning of Comma, we published a paper called Learning a Driving Simulator.

    14. LF

      Mm-hmm.

    15. GH

      And the way this thing worked was it's a, it was an autoencoder and then an RNN in the middle, right? Uh, you take an autoencoder, you compress the picture, you use an RNN, predict the next state, and these things work. You know, it was a laughably bad simulator, right? This is 2015 era machine learning technology. Today, we have VQ-VAE and transformers.

    16. LF

      Mm-hmm.

    17. GH

      We're building DriveGPT, basically.

    18. LF

      DriveGPT. Okay. Uh, so and it's trained on what? Is it trained in a self-supervised way?

    19. GH

      Yeah. It's trained on all the driving data to predict the next frame.

    20. LF

      So really trying to, uh, learn a human policy.

    21. GH

      Yeah.

    22. LF

      What would a human do?

    23. GH

      Well, actually our simulator is conditioned on the pose. So it's, it's actually a simulator. You can put in like a state action pair and get out the next state.

    24. LF

      Okay.

    25. GH

      Um, and then once you have a simulator, you can do RL in the simulator, and RL will get us that human policy.

    26. LF

      So it transfers?

    27. GH

      Yeah. RL with a reward function, not asking is this close to the human policy? But asking would a human disengage if you did this behavior?

    28. LF

      Okay. Let me think about the, the distinction there. Would a human disengage?

    29. GH

      Yeah.

    30. LF

      Would a human disengage? That, um, correlates, I guess, with human policy, but it could be different. So it's, it, uh, it doesn't just say, "What would a human do?" It says, "What would a good human driver do-"

Episode duration: 3:08:45

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode dNrTrx42DGQ

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome