George Hotz: Comma.ai, OpenPilot, and Autonomous Vehicles | Lex Fridman Podcast #31

1. 0:00 – 3:55

   Simulation theory, virtual machines, and formal proofs of “correct” worlds

   1. LFLex Fridman0:00

      The following is a conversation with George Hotz. He's the founder of Comma.ai, a machine learning based vehicle automation company. He is most certainly an outspoken personality in the field of AI and technology in general. He first gained recognition for being the first person to carrier unlock an iPhone and since then, he's done quite a few interesting things at the intersection of hardware and software. This is the Artificial Intelligence podcast. If you enjoy it, subscribe on YouTube, give it five stars on iTunes, support it on Patreon, or simply connect with me on Twitter @lexfridman, spelled F-R-I-D-M-A-N. And I'd like to give a special thank you to Jennifer from Canada for her support of the podcast on Patreon. Merci beaucoup, Jennifer. She's, uh, been a friend and an engineering colleague for many years since I was in grad school. Your support means a lot and inspires me to keep this series going. And now here's my conversation with George Hotz. Do you think we're living in a simulation?

   2. GHGeorge Hotz1:06

      Y- yes, but it may be unfalsifiable.

   3. LFLex Fridman1:10

      What do you mean by unfalsifiable?

   4. GHGeorge Hotz1:12

      So if the simulation is designed in such a way that they did like a formal proof to show that no information can get in and out, and if their hardware is designed to, for the, anything in the simulation to always keep the hardware in spec, it may be impossible to prove whether we're in a simulation or not.

   5. LFLex Fridman1:32

      So they've designed it such that it's a closed system, you can't get outside the system?

   6. GHGeorge Hotz1:37

      Well, maybe it's one of three worlds. We're either in a simulation which can be exploited, we're in a simulation which not only can't be exploited but like the same thing's true about VMs. Um, a really well-designed VM, you can't even detect if you're in a VM or not.

   7. LFLex Fridman1:50

      (laughs) That's brilliant. So we're, uh, it's, yeah, so the simulation's running on a, on a virtual machine?

   8. GHGeorge Hotz1:56

      Yeah. But now i- in reality all VMs have ways to detect.

   9. LFLex Fridman1:59

      That's the point. I mean, is it, uh, y- you've done quite a bit of hacking yourself, uh, and so you should know that, uh, really any complicated system will have ways in and out.

   10. GHGeorge Hotz2:11

       So this isn't necessarily true going forward. I spent my time away from Comma, I learned, uh, Coq.

   11. LFLex Fridman2:19

       Mm-hmm.

   12. GHGeorge Hotz2:19

       It's a dependently typed, like, uh, it's a language we're writing math proofs in.

   13. LFLex Fridman2:23

       Mm-hmm.

   14. GHGeorge Hotz2:24

       And if you write code that compiles in a language like that, it is correct by definition. The, the types check its correctness.

   15. LFLex Fridman2:32

       Mm-hmm.

   16. GHGeorge Hotz2:33

       So it's possible that the simulation is written in a language like this, in which case, you know...

   17. LFLex Fridman2:39

       Yeah, but that, that can't be sufficiently expressive a language like that.

   18. GHGeorge Hotz2:43

       Oh, it can.

   19. LFLex Fridman2:44

       It can be?

   20. GHGeorge Hotz2:45

       Oh, yeah.

   21. LFLex Fridman2:46

       Okay. Well so you, uh, hm, all right, so-

   22. GHGeorge Hotz2:48

       The simulation doesn't have to be Turing-complete if it has a scheduled end date. Looks like it does actually with Entropy.

   23. LFLex Fridman2:54

       I, uh, I don't think that, uh, a simulation as, that results in, uh, something as complicated as the universe wo- would have a form of proof th- of correctness, right? Uh, it's possible of course, uh...

   24. GHGeorge Hotz3:10

       We have no idea how good their tooling is, and we have-

   25. LFLex Fridman3:13

       True.

   26. GHGeorge Hotz3:13

       ... no idea how complicated the universe computer really is. It may be quite simple.

   27. LFLex Fridman3:18

       It's just very large, right?

   28. GHGeorge Hotz3:19

       It's very, it's definitely very large.

   29. LFLex Fridman3:22

       But the fundamental rules might be super simple.

   30. GHGeorge Hotz3:24

       Yeah. Conway's game of life kind of stuff.
2. 3:55 – 6:32

   “Thinking upwards”: new narratives, VR as a desirable reality, and non-zero-sum gratitude

   1. GHGeorge Hotz3:55

      It wasn't a very practical talk about how to actually escape a simulation. It was more about a way of restructuring, uh, an us versus them narrative. If we continue on the path we're going with technology, I think we're in big trouble, like as a species, and not just as a species but even as me as an individual member of the species. So if we could change rhetoric to be more like to think upwards, like to think about that we're in a simulation and how we could get out, already we'd be on the right path. What you actually do once you do that, well, I assume I would have acquired way more intelligence in the process of doing that, so I'll just ask that.

   2. LFLex Fridman4:39

      So the th- the thinking upwards, what kind of ideas, what kind of breakthrough ideas do you think thinking in that way could ins- inspire? And why did you say upwards?

   3. GHGeorge Hotz4:49

      Upwards-

   4. LFLex Fridman4:50

      Into space? Are you thinking sort of exploration in all forms?

   5. GHGeorge Hotz4:54

      The space narrative, uh, that held for the modernist generation doesn't hold as well for the postmodern generation.

   6. LFLex Fridman5:04

      What's the space narrative? Are we talking about the same space? The three-dimensional space?

   7. GHGeorge Hotz5:07

      Yeah. No, no, no, space-

   8. LFLex Fridman5:07

      Like literally space, yeah.

   9. GHGeorge Hotz5:09

      ... like going up in space. Like building, like Elon Musk, like we're gonna build rockets, we're gonna go to Mars, we're gonna colonize the universe.

   10. LFLex Fridman5:13

       And the narrative you're referring... I, I was born in the Soviet Union.

   11. GHGeorge Hotz5:16

       Sure.

   12. LFLex Fridman5:16

       You're referring to the race to space-

   13. GHGeorge Hotz5:18

       The race to space, yes.

   14. LFLex Fridman5:18

       ... ex- explore, okay.

   15. GHGeorge Hotz5:19

       That was a great modernist narrative.

   16. LFLex Fridman5:21

       Yeah.

   17. GHGeorge Hotz5:23

       It doesn't seem to hold the same weight in today's culture.

   18. LFLex Fridman5:27

       Hmm.

   19. GHGeorge Hotz5:27

       I'm hoping for good postmodern narratives that replace it.

   20. LFLex Fridman5:32

       So think, let's think, so you work a lot with AI, so AI is one formulation of that narrative. There could be also, I don't know how much you do in VR and AR.

   21. GHGeorge Hotz5:42

       Yeah.

   22. LFLex Fridman5:42

       That's another... I, I know less about it but every time I play with it and, and I research, it's fascinating that virtual world. Are you, are you interested in the virtual world?

   23. GHGeorge Hotz5:52

       I would like to move to virtual reality.

   24. LFLex Fridman5:55

       Uh, in terms of your work or in terms-

   25. GHGeorge Hotz5:56

       No, I would like to physically move there. The apartment I can rent in the cloud is way better than the apartment I can rent in the real world.

   26. LFLex Fridman6:03

       Well, it's all relative, isn't it? Because others will have very nice apartments too, so you'll be inferior in the virtual world as well.

   27. GHGeorge Hotz6:09

       No, that's not how I view the world, right? I don't view the world... I mean, that's a very like, like almost zero summish way to view the world, say like, my great apartment isn't great because my neighbor has one too.

   28. LFLex Fridman6:20

       Right.

   29. GHGeorge Hotz6:20

       No, my great apartment is great because, like, look at this dishwasher, man.

   30. LFLex Fridman6:24

       Yeah.
3. 6:32 – 8:45

   From iPhone unlock to hacker mindset: hardware-first origins and learning software exploitation

   1. LFLex Fridman6:32

      Yeah. So you have fundamental gratitude. The- the world first learned of, uh, Geohot, George Hotz, in August 2007, maybe before then, but certainly in August 2007, when you were the first person to, uh, unlock, carry unlock an iPhone.

   2. GHGeorge Hotz6:48

      Mm-hmm.

   3. LFLex Fridman6:48

      How did you get into hacking? What was the first system you discovered vulnerabilities for and broke into?

   4. GHGeorge Hotz6:56

      So, that was really kind of the first thing.

   5. LFLex Fridman7:01

      Really?

   6. GHGeorge Hotz7:01

      I had a- I had a book in- in 2006 called Gray Hat Hacking, and I guess I realized that if you acquired these sort of powers, you could control the world.

   7. LFLex Fridman7:15

      Mm-hmm.

   8. GHGeorge Hotz7:16

      But I didn't really know that much about computers back then. I started with electronics. The first iPhone hack was physical.

   9. LFLex Fridman7:24

      Hardware, yeah.

   10. GHGeorge Hotz7:24

       Um, you had to open it up and pull an address line high. And it was because I didn't really know about software exploitation. I learned that all in the next few years and I got very good at it. But back then, I knew about like how memory chips are connected to processors and stuff.

   11. LFLex Fridman7:39

       You knew about software and programming, you just didn't- didn't know-

   12. GHGeorge Hotz7:42

       No.

   13. LFLex Fridman7:43

       Oh, really?

   14. GHGeorge Hotz7:43

       No.

   15. LFLex Fridman7:43

       So your- (laughs) your view of the world and computers was, uh, physical-

   16. GHGeorge Hotz7:48

       Yeah.

   17. LFLex Fridman7:48

       ... was com- was hardware.

   18. GHGeorge Hotz7:49

       Actually, if you read the code that I released with that-

   19. LFLex Fridman7:52

       Mm-hmm.

   20. GHGeorge Hotz7:52

       ... in August 2000, uh, 7, it's atrocious.

   21. LFLex Fridman7:55

       What language was it?

   22. GHGeorge Hotz7:56

       C.

   23. LFLex Fridman7:57

       C, nice.

   24. GHGeorge Hotz7:58

       And in a broken sort of state machine-esque C. I- I didn't know how to program.

   25. LFLex Fridman8:02

       Yeah. So how did you learn to program? What was your journey? 'Cause, I mean, we'll talk about it. You've, uh, live streamed some of your programming.

   26. GHGeorge Hotz8:12

       Yeah.

   27. LFLex Fridman8:12

       This, uh, chaotic, beautiful mess. How did you arrive at that?

   28. GHGeorge Hotz8:16

       Years and years of practice. I interned at Google, uh, after- the summer after the iPhone unlock.

   29. LFLex Fridman8:24

       Hmm.

   30. GHGeorge Hotz8:24

       And I did a contract for them where I built hardware for- for street view and I wrote a software library to interact with it.
4. 8:45 – 14:05

   Kira: building a “timeless debugger” and what CTFs teach about breaking systems

   1. GHGeorge Hotz8:45

      The way that I really got good was I wanted to write this thing, like, that could, uh, emulate and then, uh, visualize like arm binaries, 'cause I wanted to hack the iPhone better.

   2. LFLex Fridman9:00

      Mm-hmm.

   3. GHGeorge Hotz9:00

      And I didn't like that I couldn't like see what the... I couldn't single step through the processor because I had no debugger on there, especially for the low level things like the boot ROM and the bootloader. So I tried to build this tool to do it. And I built the tool once, and it was terrible. I built the tool a second time, it was terrible. I built the tool a third time, this was by the time I was at Facebook, it was kind of okay. And then I built the tool a fourth time when I was a Google intern again in 2014, and that was the first time I was like, "This is finally usable."

   4. LFLex Fridman9:25

      How do you pronounce this, Kira?

   5. GHGeorge Hotz9:27

      Kira, yeah.

   6. LFLex Fridman9:28

      So it's essentially the most efficient way to visualize the change of state of the computer as the program is running. That's what you mean by debugger, right?

   7. GHGeorge Hotz9:39

      Yeah, it's a timeless debugger so you can rewind just as easily as going forward. Think about if you're using GDB, you have to put a watch on a variable-

   8. LFLex Fridman9:47

      Mm-hmm.

   9. GHGeorge Hotz9:47

      ... if you want to see if that variable changes. In Kira, you can just click on that variable and then it shows every single time when that variable was changed or accessed. Think about it like Git for your computer's, uh, the- the run log.

   10. LFLex Fridman9:59

       So there's like a- a deep log of, uh, of the state of the computer as the program runs and you can rewind. Why isn't that... or maybe it is, maybe you can educate me, why isn't that kind of debugging used more often?

   11. GHGeorge Hotz10:14

       Uh, 'cause the tooling's bad. Uh, well, two things. One, if you're trying to debug Chrome, Chrome is a 200 megabyte binary that runs slowly on desktops.

   12. LFLex Fridman10:25

       Mm-hmm.

   13. GHGeorge Hotz10:25

       So that's gonna be really hard to use for that. But it's really good to use for like CTFs and for boot ROMs and for small parts of code.

   14. LFLex Fridman10:33

       Mm-hmm.

   15. GHGeorge Hotz10:33

       So it's- it's hard if you're trying to debug like massive systems.

   16. LFLex Fridman10:36

       What's a CTF and what's a boot ROM?

   17. GHGeorge Hotz10:38

       A boot ROM is the first code that executes the minute you give power to your iPhone.

   18. LFLex Fridman10:42

       Okay.

   19. GHGeorge Hotz10:43

       And CTF were these competitions that I played, Capture the Flag.

   20. LFLex Fridman10:46

       Capture the Flag. I was gonna ask you about that.

   21. GHGeorge Hotz10:48

       Yeah.

   22. LFLex Fridman10:48

       What are those? Those look at... I watched a couple of videos on YouTube, those look fascinating. What have you learned about maybe at the high level of vulnerability of systems from these competitions?

   23. GHGeorge Hotz10:59

       The, like... I feel like, like in the heyday of CTFs, you had all of the best security people in the world-

   24. LFLex Fridman11:07

       Mm-hmm.

   25. GHGeorge Hotz11:08

       ... challenging each other-

   26. LFLex Fridman11:09

       Yeah, that's-

   27. GHGeorge Hotz11:10

       ... and coming up with new toy exploitable things over here and then everybody, "Okay, who can break it?" And when you, when you break it, you get like... there's like a file in the server called Flag.

   28. LFLex Fridman11:18

       Mm-hmm.

   29. GHGeorge Hotz11:19

       And then there's a program running listening on a socket that's vulnerable. So you write an exploit, you get a shell, and then you cat Flag, and then you type the flag into like a web-based scoreboard and you get points.

   30. LFLex Fridman11:29

       So the goal is essentially to find an exploit in the system that allows you to run Shell, to run arbitrary code on that system.
5. 14:05 – 20:15

   Security ethics and incentives: why not crime, why defense is hard, and “trust what it does”

   1. GHGeorge Hotz14:05

      Oh yeah, I mean, doing criminal stuff is not really worth it. And I'll actually use the same argument for why I don't do defense for why I don't do crime. If you want to defend a system, say the system has 10 holes, right? If you find nine of those holes as a defender, you still lose because the attacker gets in through the last one. If you're an attacker, you only have to find one out of the 10. But if you're a criminal, if you log on with a VPN nine out of the 10 times, but one time you forget, you're done.

   2. LFLex Fridman14:37

      Because you're caught. Okay. I-

   3. GHGeorge Hotz14:39

      Because you only have to mess up once to be caught as a criminal.

   4. LFLex Fridman14:42

      Yeah.

   5. GHGeorge Hotz14:42

      That's why I'm not a criminal.

   6. LFLex Fridman14:44

      (laughs) But, okay, let me, uh... 'Cause I was having a discussion with somebody, uh, just at, at a high level about nuclear weapons actually, why we're having blown ourselves up yet. And my feeling is all the smart people in the world, if you look at the distribution of, of smart people, smart people are generally good. And then this other person, I was talking to Sean Carroll, the physicist, and he was saying, "No, good and bad people are evenly distributed amongst everybody." My sense was good hackers are in general good people and they don't want to mess with the world. What's your sense?

   7. GHGeorge Hotz15:21

      I'm not even sure about that. Like, I have a nice life. Crime wouldn't get me anything. But if you're good and you have these skills, you probably have a nice life too, right? Like-

   8. LFLex Fridman15:38

      Right, you can use to further things. But is there an ethical... Is there some... Is there a little voice in your head that says, uh, "Well, yeah, if you could hack something to where you could hurt people and you could earn a lot of money doing it, though. Not hurt physically perhaps, but disrupt their life in some kind of way." I- isn't there a little voice that says-

   9. GHGeorge Hotz16:02

      Um, well, two things. One, I don't really care about money.

   10. LFLex Fridman16:06

       Right.

   11. GHGeorge Hotz16:06

       So like the money wouldn't be an incentive. The thrill might be an incentive.

   12. LFLex Fridman16:10

       The thrill.

   13. GHGeorge Hotz16:10

       But when I was 19, I read Crime and Punishment.

   14. LFLex Fridman16:13

       Right. Good.

   15. GHGeorge Hotz16:14

       That was another, that was another great one that talked me out of ever really doing crime.

   16. LFLex Fridman16:18

       (laughs)

   17. GHGeorge Hotz16:19

       Um, 'cause it's like, "That's gonna be me. I'd get away with it," but it would just run through my head. Even if I got away with it, you know? And then you do crime for long enough, you'll never get away with it.

   18. LFLex Fridman16:28

       That's right, in the end. That's a good reason to be good.

   19. GHGeorge Hotz16:32

       I wouldn't say I'm good.

   20. LFLex Fridman16:33

       (laughs)

   21. GHGeorge Hotz16:33

       I would just say I'm not bad.

   22. LFLex Fridman16:35

       You're a talented programmer and a, a hacker in a good positive sense of the word, word. You've, uh, played around, found vulnerabilities in various systems. What have you learned broadly about the design of systems and so on from that, from that whole process?

   23. GHGeorge Hotz16:53

       You learn to not take things for what people say they are.

   24. LFLex Fridman17:02

       Mm-hmm.

   25. GHGeorge Hotz17:02

       But you look at things for what they actually are.

   26. LFLex Fridman17:06

       Hmm. Yeah.

   27. GHGeorge Hotz17:08

       I understand that's what you tell me it is, but what does it do?

   28. LFLex Fridman17:11

       Right. And you have nice visualization tools to really know what it's really doing.

   29. GHGeorge Hotz17:16

       Oh, I wish... I'm a better programmer now than I was in 2014. I said, "Cura, that was the first tool that I wrote that was usable." I wouldn't say the code was great. I still wouldn't say my code is great. It's better.

   30. LFLex Fridman17:28

       So how was your evolution as a programmer except practice? You went... You started with C, at which point did you pick up Python? 'Cause you're pretty big in Python now.
6. 20:15 – 26:49

   Programming style, learning, and language tradeoffs: Python vs Go, and why JS ecosystems hurt

   1. LFLex Fridman20:15

      I watched your coding sessions on, that you streamed online.

   2. GHGeorge Hotz20:18

      Yeah.

   3. LFLex Fridman20:19

      Uh, you code things up, basic projects usually from scratch. I would say, sort of as a programmer myself just watching you, that you type really fast and your brain works in both brilliant and chaotic ways. Uh, I don't know if that's always true, but certainly for the live streams. So it's interesting to me because I'm more, I'm much slower and systematic and careful, and you just move, I mean, probably in an order of magnitude faster. So I'm curious, is there a method to your madness? Is, is it some- or is it just who you are?

   4. GHGeorge Hotz20:53

      There's pros and cons. Um, there's pros and cons to my programming style, and I'm aware of them. Like, if, if you ask me to, like, like get something up and working quickly with like an API that's kind of undocumented-

   5. LFLex Fridman21:06

      Yeah.

   6. GHGeorge Hotz21:06

      ... I will do this super fast 'cause I will throw things at it until it works. If you ask me to take a vector and rotate it 90 degrees and then flip it over the XY plane (laughs) , I'll spam program for two hours and won't get it.

   7. LFLex Fridman21:22

      Oh, because it's something that you can just do with a sheet of paper, think through-

   8. GHGeorge Hotz21:25

      Yeah.

   9. LFLex Fridman21:25

      ... design, and then just, uh, so you really just throw stuff at the wall and you get so good at it that it usually works.

   10. GHGeorge Hotz21:34

       I should become better at the other kind as well. Sometimes I'll do things methodically. It's nowhere near as entertaining on the Twitch streams. I do exaggerate it a bit on-

   11. LFLex Fridman21:42

       Yeah.

   12. GHGeorge Hotz21:42

       ... the Twitch streams as well. The Twitch streams, I mean, what do you want to see a gamer, you want to see actions per minute, right?

   13. LFLex Fridman21:46

       Yeah. Yeah.

   14. GHGeorge Hotz21:46

       Well, I'll show you APM for programming too.

   15. LFLex Fridman21:48

       Yeah. I recommend people go to it. I think I watched, I watched probably several hours of you pro- like I've actually left you programming in the background-

   16. GHGeorge Hotz21:57

       (laughs)

   17. LFLex Fridman21:57

       ... while I was programming-

   18. GHGeorge Hotz21:59

       (laughs)

   19. LFLex Fridman21:59

       ... because you made me, you're, it was, it was like wa- watching a really good gamer, it's like energizes you 'cause you're like moving so fast and so it's, it's awesome, it's inspiring and it's almo- it made me jealous that like, because my own programming is inadequate in terms of speed-

   20. GHGeorge Hotz22:15

       Oh, I-

   21. LFLex Fridman22:15

       ... so I was like... (laughs)

   22. GHGeorge Hotz22:17

       So I'm, I'm, I'm, I'm twice as frantic on the live streams-

   23. LFLex Fridman22:20

       Yeah.

   24. GHGeorge Hotz22:20

       ... as I am when I code without-

   25. LFLex Fridman22:22

       Right. Well, it's-

   26. GHGeorge Hotz22:22

       ... uh...

   27. LFLex Fridman22:22

       ... super entertaining so I, I wasn't even paying attention to what you were coding, which is great. (laughs) It's just watching you switch windows and, uh, Vim I guess is the most

   28. GHGeorge Hotz22:31

       Yeah. Yeah, this is Vim and screen. I've developed a workflow at Facebook and stuck with it.

   29. LFLex Fridman22:35

       How do you learn new programming tools, ideas, techniques these days? What's your, like, uh, methodology for learning new things?

   30. GHGeorge Hotz22:42

       So I wrote for Komma, the distributed file systems out in the world are extremely complex. Like if you want to install something like, like, like Ceph, uh, Ceph is I think the, like, open infrastructure, uh, distributive file system, or there's like newer ones like SeaweedFS, but these are all like 10,000 plus line projects. I think some of them are even 100,000 line, and just configuring them is a nightmare. So I wrote, uh, I wrote one. Um, it's 200 lines and it's, it uses like NGINX for the live servers and has this little master server that I wrote in Go. And the way I-
7. 26:49 – 32:19

   Founding Comma.ai: the Tesla/Elon contract story and the reality of building AP1-level capability

   1. LFLex Fridman26:49

      Uh, you founded Comma AI. Let's, uh, at a high level, how did you get into the world of vehicle automation? Can you also just, for people who don't know, tell the story of Comma.AI?

   2. GHGeorge Hotz27:01

      Sure. So I was working at this AI startup and, uh, a friend approached me and he's like, "Dude, I don't know where this is going, but the coolest applied AI problem today is self-driving cars."

   3. LFLex Fridman27:16

      Yep.

   4. GHGeorge Hotz27:16

      I'm like, "Well, absolutely." "Do you want to meet with, uh, Elon Musk?"

   5. LFLex Fridman27:20

      Hmm.

   6. GHGeorge Hotz27:20

      "And, uh, uh, he's looking for somebody to build a vision system-"

   7. LFLex Fridman27:24

      Mm-hmm.

   8. GHGeorge Hotz27:24

      "... for, uh, autopilot." This is when they were still on AP1, they were still using Mobileye.

   9. LFLex Fridman27:30

      Yep.

   10. GHGeorge Hotz27:30

       Elon, back then, was looking for a replacement and he brought me in and we talked about a contract where I would deliver something that meets Mobileye-level performance, uh, I would get paid $12 million if I could deliver it tomorrow and I would lose $1 million for every month I didn't deliver.

   11. LFLex Fridman27:46

       Yeah.

   12. GHGeorge Hotz27:47

       Um, so I was like, "Okay, this is great deal, this is a super exciting challenge. You know what? Even if it takes me 10 months, I get $2 million, it's good. Maybe I can finish up in five, maybe I don't finish it all and I get paid nothing and I'll work for 12 months for free."

   13. LFLex Fridman28:00

       So maybe, uh, just take a pause on that. I'm also curious about this because I've been working in robotics for a long time and I'm curious to see a person like you just step in and sort of, um, somewhat naive, but brilliant, right? So that's th- though that's the best place to be 'cause you basically full steam take on a problem. How confident, how, from that time, 'cause you know a lot more now, at that time, how hard do you think it is to solve all of autonomous driving?

   14. GHGeorge Hotz28:26

       I remember I suggested to Elon in the meeting, um, putting a GPU behind each camera to keep the compute local. This is an incredibly stupid idea. I leave the meeting 10 minutes later and I'm like, "I could have spent a little bit of time thinking about this problem before I would've-"

   15. LFLex Fridman28:42

       Why is this a stupid idea?

   16. GHGeorge Hotz28:44

       Oh, just send all your cameras to one big GPU, you're much better off doing that.

   17. LFLex Fridman28:48

       Oh, sorry, you said behind every camera have a GPU.

   18. GHGeorge Hotz28:50

       Every camera have a small GPU. I was like, "Oh, I'll put the first few layers of my comp there." Ugh. Like, why'd I say that?

   19. LFLex Fridman28:56

       That's possible. I mean-

   20. GHGeorge Hotz28:57

       It's possible, but it's a bad idea. It's, it's-

   21. LFLex Fridman28:59

       Uh, it's not obviously a bad idea.

   22. GHGeorge Hotz29:00

       Pretty obviously a bad... But whether it's actually a bad idea or not, I, I left that meeting with Elon like beating myself up. I'm like, "Why'd I say something stupid?"

   23. LFLex Fridman29:07

       Yeah, you haven't, like you haven't-

   24. GHGeorge Hotz29:08

       Um-

   25. LFLex Fridman29:08

       ... at least like thought through every aspect fully, yeah.

   26. GHGeorge Hotz29:12

       Well, he's very sharp, too.

   27. LFLex Fridman29:13

       Yeah.

   28. GHGeorge Hotz29:13

       Like usually in life, I get away with saying stupid things and then kind of course... Oh, right, right away he called me out about it. And like usually in life I get away with saying stupid things-

   29. LFLex Fridman29:21

       Yeah.

   30. GHGeorge Hotz29:21

       ... and then like people will, uh, you know, pe- a lot of times people don't even notice and I'll like correct it and bring the conversation back.
8. 32:19 – 35:46

   Lane keeping as the real near-term value: critique of Navigate on Autopilot and Waymo’s value proposition

   1. GHGeorge Hotz32:19

      I think we've kept pace with them pretty well. I think Navigator and autopilot is terrible. We had some demo features internally of the same stuff, and we would test it, and I'm like, "I'm not shipping this even as, like, open source software to people."

   2. LFLex Fridman32:35

      Why do you think it's terrible?

   3. GHGeorge Hotz32:37

      Consumer Reports does a great job of describing it. Like, when it makes a lane change, it does it worse than a human.

   4. LFLex Fridman32:43

      Mm-hmm.

   5. GHGeorge Hotz32:43

      You shouldn't ship things like... Autopilot, OpenPilot, they lane-keep better than a human.

   6. LFLex Fridman32:49

      Right.

   7. GHGeorge Hotz32:49

      If you turn it on for a stretch of, uh, highway, like an hour long, it's never gonna touch a lane line. A human will touch probably a lane line twice.

   8. LFLex Fridman32:59

      You just inspired me. I don't know if y- you're grounded in data on that. But-

   9. GHGeorge Hotz33:02

      I, I read your paper.

   10. LFLex Fridman33:03

       Okay.

   11. GHGeorge Hotz33:03

       (laughs)

   12. LFLex Fridman33:03

       But that... No, but that's interesting. Uh, I wonder actually how often we touch lane lines, uh, in gen- like, a l- a little bit, 'cause it is the-

   13. GHGeorge Hotz33:13

       I could, I could answer that question pretty easily with the Comma dataset.

   14. LFLex Fridman33:15

       Yeah. I'm curious. I, I-

   15. GHGeorge Hotz33:17

       I, I've never answered it. I don't know.

   16. LFLex Fridman33:18

       Yeah, yeah.

   17. GHGeorge Hotz33:18

       I, I just... two is like my personal-

   18. LFLex Fridman33:19

       It's, it feels-

   19. GHGeorge Hotz33:20

       Yeah.

   20. LFLex Fridman33:20

       ... it feels right. Well, that's interesting 'cause that every time you touch a lane, that's a source of, um, a little bit of stress, and kind of lane keeping is removing that stress.

   21. GHGeorge Hotz33:29

       Yeah.

   22. LFLex Fridman33:29

       That's all to me the big, the biggest value-add, honestly, is just removing the stress of having to stay in lane. And I think honestly, I don't think people fully realize first of all that that's a big value-add, uh, but also that that's all it is.

   23. GHGeorge Hotz33:45

       And that... Not only I find it a huge value-add. I, I drove down... when we moved to San Diego, I drove down in o- Enterprise rent-a-car-

   24. LFLex Fridman33:52

       Mm-hmm.

   25. GHGeorge Hotz33:52

       ... and I missed it so... I missed having the system so much. It, it, it's so much more tiring-

   26. LFLex Fridman33:58

       Yeah.

   27. GHGeorge Hotz33:58

       ... to drive without it. It's, it is that lane centering that's the key feature.

   28. LFLex Fridman34:04

       Yeah.

   29. GHGeorge Hotz34:06

       And in a way, it's the only feature that actually adds value to people's lives in autonomous vehicles today. Waymo does not add value to people's lives. It's a more expensive, slower, slower Uber. Maybe someday it'll be this big cliff where it adds value, but-

   30. LFLex Fridman34:18

       Yes.
9. 35:46 – 41:45

   OpenPilot hardware + integration: “a phone in a plastic case,” CAN proxying, and supported vehicles

   1. GHGeorge Hotz35:46

      So OpenPilot as the, the, the box that we sell that it runs on, it's a phone in a plastic case.

   2. LFLex Fridman35:53

      Mm-hmm.

   3. GHGeorge Hotz35:54

      Um, it's nothing special. We sell it without the software. So you're like, you know, you buy the phone and it's just easy, it'll be easy set up, but it's sold with no software. OpenPilot right now is about to be 0.6. When it gets to 1.0, I think we'll be ready for a consumer product. We're not gonna add any new features. We're just gonna make the lane keeping really, really good.

   4. LFLex Fridman36:14

      Okay, I gotcha then. (laughs)

   5. GHGeorge Hotz36:15

      Um, so what do we have right now?

   6. LFLex Fridman36:16

      Yeah.

   7. GHGeorge Hotz36:16

      Uh, it's a, uh, Snapdragon 820. It's a Sony IMX298 forward-facing camera.

   8. LFLex Fridman36:24

      Mm-hmm.

   9. GHGeorge Hotz36:25

      Um, driver monitoring camera, which is a selfie cam on the phone. And, uh...... a CAN transceiver, maybe, there's a thing called pandas.

   10. LFLex Fridman36:33

       Mm-hmm.

   11. GHGeorge Hotz36:33

       And they talk over USB to the phone, and then they have three CAN buses that they talk to the car on. Uh, one of those CAN buses is the radar CAN bus, one of them is the main car CAN bus, and the other one is the proxy camera CAN bus.

   12. LFLex Fridman36:46

       Mm-hmm.

   13. GHGeorge Hotz36:46

       We leave the existing camera in place so we don't turn AEB off. Uh, right now, we still turn-

   14. LFLex Fridman36:51

       Yeah.

   15. GHGeorge Hotz36:52

       ... AEB off if you're using our longitudinal, but we're going to fix that before 1.0.

   16. LFLex Fridman36:55

       Got it. Wow. That's cool. So, and it's CAN both ways. So, how are you able to control vehicles?

   17. GHGeorge Hotz37:03

       So, we proxy... The vehicles that we work with already have a lane keeping assist system.

   18. LFLex Fridman37:09

       Mm-hmm.

   19. GHGeorge Hotz37:10

       So, lane keeping assist can mean a huge variety of things.

   20. LFLex Fridman37:13

       Okay.

   21. GHGeorge Hotz37:13

       It can mean... It will apply a small torque to the wheel after you've already crossed a lane line by a foot-

   22. LFLex Fridman37:20

       Mm-hmm.

   23. GHGeorge Hotz37:21

       ... which is the system in the older Toyotas. Versus, like, I think Tesla still calls it lane keeping assist, where it'll keep you perfectly in the center of the lane, uh, on the highway.

   24. LFLex Fridman37:32

       You can control, like you-

   25. GHGeorge Hotz37:33

       Yeah.

   26. LFLex Fridman37:33

       ... with a joystick, the car. These, so these cars already have the capability of drive-by-wire. So, uh, is it... Is it trivial to convert a car that it operates with, uh, it c- it, uh, OpenPilot's able to control the steering?

   27. GHGeorge Hotz37:47

       Yeah. Oh, a new car or a car that we... So we have support now for 45 different makes of cars.

   28. LFLex Fridman37:52

       What are, what other cars in general?

   29. GHGeorge Hotz37:54

       Uh, mostly Hondas and Toyotas. We support almost every Honda and Toyota made, uh, this year. And then a bunch of GMs, bunch of Subarus, bunch of Chevys.

   30. LFLex Fridman38:04

       So it doesn't have to be like a Prius, it could be Corolla as well, and that kind of...
10. 41:45 – 49:35

    Driver monitoring and safety UX: cheat-proof attention checks, low-friction takeover, and torque limits

    1. GHGeorge Hotz41:45

       Two things.

    2. LFLex Fridman41:45

       ... for the safety of the drivers?

    3. GHGeorge Hotz41:47

       One, we don't consider OpenPilot to be 1.0 until we have 100% driver monitoring. You, you can cheat right now our driver monitoring system. There's a few ways to cheat it. They're pretty obvious. Um, we're working on making that better. Before we ship a consumer product that can drive cars, I want to make sure that I have driver monitoring that you can't cheat.

    4. LFLex Fridman42:05

       What's, like, a successful driver monitoring system look like? It's keep... It's, is it all about just keeping your eyes on the road?

    5. GHGeorge Hotz42:11

       Um, well, a few things. So that's what we went with at first for driver monitoring. I'm checking, I'm actually looking at where your head is looking. The camera's not that high resolution; eyes are a little bit hard to get.

    6. LFLex Fridman42:21

       Well, head is, is big. I mean, that's, uh...

    7. GHGeorge Hotz42:23

       Head is, head is good. And actually a lot of it, just v- uh, psychology wise, to have that monitor constantly there, it reminds you that you have to be paying attention. But we want to go further. We just hired someone full time to come on and do the driver monitoring. Uh, I want to detect phone in frame.... and I want to make sure you're not sleeping.

    8. LFLex Fridman42:42

       How much does the camera see of the body?

    9. GHGeorge Hotz42:45

       This one, not enough.

    10. LFLex Fridman42:47

        Not enough.

    11. GHGeorge Hotz42:48

        The next one, everything.

    12. LFLex Fridman42:50

        Well, it's interesting now, fisheye, 'cause we have, uh, we're doing just data collection, not real time.

    13. GHGeorge Hotz42:55

        Yeah.

    14. LFLex Fridman42:55

        But fisheye is a beautiful, uh-

    15. GHGeorge Hotz42:57

        Yeah.

    16. LFLex Fridman42:57

        ... being able to capture the body, and the smartphone is really like the biggest problem.

    17. GHGeorge Hotz43:03

        I'll, I'll show you, uh, I can show you one of the pictures from, from our, our new system.

    18. LFLex Fridman43:07

        Awesome. So you're basically saying the driver monitoring will be the answer to that.

    19. GHGeorge Hotz43:12

        Um, I think the other, uh, point that, that you raised in your paper is, is good as well. You're, you're not asking a human to, uh, supervise a machine, uh, without giving them the... they can take over at any time.

    20. LFLex Fridman43:23

        Right. There's-

    21. GHGeorge Hotz43:24

        Our, our safety model, you can take over. We, we disengage on both the gas or the brake. Uh, we don't disengage on steering. I don't feel you have to. But, uh, we disengage on gas or brake. So it's very easy for you to take over.

    22. LFLex Fridman43:34

        Right.

    23. GHGeorge Hotz43:34

        And it's very easy for you to reengage. That switching should be super cheap.

    24. LFLex Fridman43:39

        Yep.

    25. GHGeorge Hotz43:39

        The cars that require... even autopilot requires a double press. That's almost... I see, I don't like that.

    26. LFLex Fridman43:44

        Yeah.

    27. GHGeorge Hotz43:44

        And then, then they cancel. Um, to cancel in autopilot, you either have to press Cancel, which no one knows where that is, so they press the brake.

    28. LFLex Fridman43:51

        Okay.

    29. GHGeorge Hotz43:51

        But a lot of times you don't actually want to press the brake.

    30. LFLex Fridman43:53

        Right.
11. 49:35 – 1:07:34

    Data-driven development: replay-based simulation, anti-mapping stance, and end-to-end perception→planning

    1. LFLex Fridman49:35

       So what are the current limitations of, uh, OpenPilot? What are the main problems that still need to be solved?

    2. GHGeorge Hotz49:40

       Um, s- we're hopefully fixing, uh, a few of them in- in- in 06. We're not as good as autopilot at stopped cars. Um, so if you're coming, uh, up to a red light at, like, 55, um, so it's the radar-stopped car problem-

    3. LFLex Fridman49:56

       Mm-hmm.

    4. GHGeorge Hotz49:57

       ... which is responsible for two autopilot accidents. Uh, it's hard to differentiate a stopped car from a, uh, like, signpost.

    5. LFLex Fridman50:03

       Yeah, a static object.

    6. GHGeorge Hotz50:05

       Um, so you have to fuse. You have to do this visually. There's no way from the radar data to tell the difference. Maybe you could make a map, but I- I don't know, I don't really believe in mapping at all anymore. Um, so-

    7. LFLex Fridman50:14

       Wait, wait, wait, what? You don't believe in mapping?

    8. GHGeorge Hotz50:16

       No.

    9. LFLex Fridman50:16

       So, you basically ... the OpenPilot solution is saying react to the environment as you see it-

    10. GHGeorge Hotz50:22

        Exactly.

    11. LFLex Fridman50:22

        ... just like human doing- beings do. Okay.

    12. GHGeorge Hotz50:24

        And then eventually when you want to do navigate on, uh, OpenPilot, I'll train the net to look at Waze. I'll run Waze in the background and I'll try and

    13. NANarrator50:31

        (laughs)

    14. LFLex Fridman50:31

        Are-

    15. GHGeorge Hotz50:31

        ... comp that on Waze.

    16. LFLex Fridman50:32

        (laughs) are you using GPS at all?

    17. GHGeorge Hotz50:34

        We use it to ground truth. We use it to very carefully ground truth the paths. Uh, we have a stack which can recover relative to 10 centimeters over one minute. Um, and then we use that to ground truth exactly where the car went in that local part of the environment, but it's all local.

    18. LFLex Fridman50:48

        How are you testing in general, just for yourself? Like, experiments and stuff? All right-

    19. GHGeorge Hotz50:52

        Um-

    20. LFLex Fridman50:53

        Uh, where are you- wh- where are you located?

    21. GHGeorge Hotz50:55

        San Diego.

    22. LFLex Fridman50:55

        San Diego?

    23. GHGeorge Hotz50:56

        Yeah.

    24. LFLex Fridman50:56

        Okay. Uh, what ... So you basically drive around there, collect some data and- and watch-

    25. GHGeorge Hotz51:02

        Um-

    26. LFLex Fridman51:02

        ... performance?

    27. GHGeorge Hotz51:03

        We have a simulator now. And we have ... our simulator's-

    28. LFLex Fridman51:05

        Nice.

    29. GHGeorge Hotz51:05

        ... really cool. Our simulator is not, uh, it's not like a unity base simulator. Our simulator lets us load in real state.

    30. LFLex Fridman51:12

        What do you mean, real estate?
12. 1:07:34 – 1:16:34

    The long game: Level 5 skepticism, business models (data + insurance), and the singularity/“winning” philosophy

    1. GHGeorge Hotz1:07:34

       It is a question. I'll say this, um, I'm here to play for 10 years. I'm not here to try to... I'm here to play for 10 years and make money along the way.

    2. LFLex Fridman1:07:42

       Mm-hmm.

    3. GHGeorge Hotz1:07:43

       I'm not here to try to promise people that I'm going to have my L5 taxi network up and working in two years.

    4. LFLex Fridman1:07:48

       Do you think that was a mistake?

    5. GHGeorge Hotz1:07:49

       Yes.

    6. LFLex Fridman1:07:50

       Uh, what do you think was the motivation behind saying that? O- other companies are also promising L5 vehicles with very different approaches in 2020, 2021-

    7. GHGeorge Hotz1:08:00

       (laughs)

    8. LFLex Fridman1:08:01

       ... 2022.

    9. GHGeorge Hotz1:08:02

       If anybody would like to bet me that those things do not pan out, I will, I will bet you. Even money? Even money, I'll bet you as much as you want. (laughs)

    10. LFLex Fridman1:08:10

        Okay. So are you worried about what's going to happen? 'Cause you and I are in full agreement on that. Like, what's going to happen when 2022, '21 come around and nobody has fleets of autonomous vehicles?

    11. GHGeorge Hotz1:08:22

        Well, you can look at the history. If you go back five years ago, they were all promised by 2018 and 2017. (laughs)

    12. LFLex Fridman1:08:30

        But they weren't that strong of promises. I mean, Ford really declared pretty... Th- I think not many have declared as f- as like definitively as they have now with these dates.

    13. GHGeorge Hotz1:08:42

        Well, okay, so let's separate L4 and L5. Do I think that it's possible for Waymo to continue to, uh, kind of like, like hack on their system until it gets to level 4 in Chandler, Arizona? Yes.

    14. LFLex Fridman1:08:55

        With no, no safety driver?

    15. GHGeorge Hotz1:08:56

        Chandler, Arizona? Yeah.

    16. LFLex Fridman1:08:59

        But by, well, sorry, which year are we talking about?

    17. GHGeorge Hotz1:09:02

        Oh, I even think that's possible by like 2020, 2021. But level 4, Chandler, Arizona, not level 5, New York City.

    18. LFLex Fridman1:09:11

        L- Level 4, meaning some very defined streets-

    19. GHGeorge Hotz1:09:15

        Yeah.

    20. LFLex Fridman1:09:16

        ... it works out really well.

    21. GHGeorge Hotz1:09:17

        Very defined streets and then practically, these streets are pretty empty. If most of the streets are covered in Waymos, Waymo can kind of change the definition of what driving is.

    22. LFLex Fridman1:09:27

        Mm-hmm.

    23. GHGeorge Hotz1:09:28

        Right? If, if, if your self-driving network is the majority of cars in an area, they only need to be safe with respect to each other and all the humans will need to learn to adapt to them. Now go drive in downtown New York. (laughs)

    24. LFLex Fridman1:09:41

        Oh, yeah, that's...

    25. GHGeorge Hotz1:09:42

        I mean, already you can talk about autonomy in like, like on farms, it already works great.

    26. LFLex Fridman1:09:47

        Mm-hmm.

    27. GHGeorge Hotz1:09:47

        Because you can really just follow the GPS line.

    28. LFLex Fridman1:09:51

        So what is, uh, success look like for Comma AI? What, um, what are the milestones, like where you can sit back with some champagne and say, "We did it, boys and girls."

    29. GHGeorge Hotz1:10:03

        Um, well, it's never over.

    30. LFLex Fridman1:10:05

        (laughs) Yeah, but-

Episode duration: 1:59:36