#147 – Dmitri Dolgov: Waymo and the Future of Self-Driving Cars

[0] The following is a conversation with Dimitri Dolgov, the CTO of Waymo, which is an autonomous driving company that started as Google self -driving car project in 2009 and became Waymo in 2016.

[1] Demetri was there all along.

[2] Waymo is currently leading in the fully autonomous vehicle space in that they actually have an at -scale deployment of publicly accessible autonomous vehicles driving passengers around with no safety driver.

[3] with nobody in the driver's seat.

[4] This, to me, is an incredible accomplishment of engineering on one of the most difficult and exciting artificial intelligence challenges of the 21st century.

[5] Quick mention of a sponsor followed by some thoughts related to the episode.

[6] Thank you to Trial Labs, a company that helps businesses apply machine learning to solve real -world problems.

[7] Blinkist, an app I use for reading through summaries of books, Better Help, Online Therapy with a licensed professional, and Cash App, the app I use to send money to friends.

[8] Please check out the sponsors in the description to get a discount at the support this podcast.

[9] As a side note, let me say that autonomous and semi -autonomous driving was the focus of my work at MIT and as a problem space that I find fascinating and full of open questions from both robotics and a human psychology perspective.

[10] There's quite a bit that I could say here about my experiences in academia on this topic that revealed to me, let's say, the less admirable size of human beings.

[11] But I choose to focus on the positive, on solutions.

[12] I'm brilliant engineers like Dmitri and the team at Waymo, who work tirelessly to innovate and to build amazing technology that will define our future.

[13] Because of Dmitri and others like him, I'm excited for this future.

[14] And who knows, perhaps I too will help contribute something of value to it.

[15] If you enjoy this thing, subscribe on YouTube, review it with five stars on upper podcast, follow on Spotify, support on Patreon, or connect with me on Twitter, and Lex Friedman.

[16] As usual, I'll do a few minutes of ads now and no ads in the middle.

[17] I try to make these interesting, but I give you timestamps, so if you skip, please still check out the sponsors by clicking the links in the description.

[18] It is, in fact, the best way to support this podcast.

[19] This episode is brought to you by Trial Labs, a new sponsor and an amazing company.

[20] They help build AI -based solutions for businesses of all sizes.

[21] I love these guys, especially after talking to them on the phone and checking out a bunch of their demos and blog posts.

[22] If you are a business or are just curious about machine learning, check them out at Trialabs .com slash Lex.

[23] they've worked on price optimization, early detection of machine failures, and all kinds of applications of computer vision, including face detection on lions.

[24] Yes, lions in support of conservation effort in Africa.

[25] Their work on price automation and optimization is probably their most impressive in terms of helping businesses make money.

[26] Also, as a cool side note, they do release open source code on GitHub on occasion like a computer vision track, for example tracking and just the general problem of occlusion very much remains unsolved but there's been a lot of exciting progress made over the past five years anyway all that to say is that trial labs is legit great engineers if you own a business and want to see how i can help you do check them out at triolabs slash Lex.

[27] That's triolabs .com slash Lex.

[28] This episode is also supported by Blinkist, my favorite app for learning new things.

[29] Blinkist takes the key ideas from thousands of nonfiction books and condenses them down into just 15 minutes that you can read or listen to.

[30] I'm a big believer in reading at least an hour every day.

[31] As part of that, I use Blinkist almost every day to try out a book I may otherwise never have a chance to read.

[32] And in general, it's a great way to broaden your view of the idea of landscape out there and find books that you may want to read more deeply.

[33] With Blinkist, you get unlimited access to read or listen to a massive library of condensed nonfiction books.

[34] I also use Blinkist shortcasts to quickly catch up on podcast episodes I've missed.

[35] Right now, Blinkist is a special offer just for the listeners of this podcast.

[36] and probably every other podcast they sponsor.

[37] But who's counting?

[38] Go to blinkus .com slash Lex to start your free seven -day trial and get 25 % off of a Blinkist premium membership.

[39] That's Blinkist spelled B -L -I -N -K -I -S -T, Blinkist .com slash Lex.

[40] To get 25 % off and a seven -day free trial, they're really making me say this over and over, aren't they?

[41] That's Blinkist .com slash Lex.

[42] This episode is also sponsored by BetterHelp, spelled H -E -L -P -Help.

[43] They figure out what you need and match you with a licensed professional therapist in under 48 hours.

[44] I chat with a person on there and enjoy it.

[45] Of course, I also have been talking to Mr. David Goggins over the past few months, who is definitely not a licensed professional therapist, but he does help me meet his and my demons and become comfortable to exist in their presence.

[46] Everyone is different, but for me, I think suffering is essential for creation, but you can suffer beautifully in a way that doesn't destroy you.

[47] Therapy can help in whatever way that therapy takes.

[48] So I think BetterHelp is an option worth trying.

[49] They're easy, private, affordable, available worldwide.

[50] You can communicate by text any time and schedule weekly audio and video sessions.

[51] You didn't ask me, but my two favorite sets, psychiatrist as Zinman Freud and Carl Jung.

[52] Their work was important in my intellectual development as a teenager.

[53] Anyway, check out BetterHelp .com slash Lex.

[54] That's betterhelp .com slash Lex.

[55] Finally, this show is presented by Cash App, the number one finance app in the App Store.

[56] When you get it, use code Lex podcast.

[57] Cash app lets you send money to friends by Bitcoin and invest in the stock market with as little as $1.

[58] I'm thinking of doing more conversations with folks who work in and around the cryptocurrency space.

[59] Similar to AI, there are a lot of charlatans in this space, but there are also a lot of freethinkers and technical geniuses that are worth exploring ideas with in depth and with care.

[60] As an example of that, Vitalik Buterin will definitely be back in the podcast.

[61] She'll listen to the first one.

[62] He'll be back on probably many more times.

[63] I see that guy accomplishing a huge amount of things in his life.

[64] I love talking to him.

[65] All right, if you get Cash App from the App Store, Google Play, use code Lex Podcast, you get $10.

[66] And Cash App will also donate $10 to First, an organization that is helping to advance robotics, STEM education for young people around the world.

[67] And now, here's my conversation with Dimitri Dolgov.

[68] When did you first fall in love with robotics or even computer science, more in general?

[69] Computer science first.

[70] at a fairly young age, the robotics happened much later.

[71] I think my first interesting introduction to computers was in the late 80s when we got our first computer.

[72] I think it was an IBM, I think IBM -AT.

[73] Do you remember those things that had like a turbo button in the front?

[74] Turbo button, the radio press it and make the thing goes faster.

[75] Did that already have floppy disks?

[76] Yeah, yeah, yeah, yeah, like the 5 .4 inch ones.

[77] I think there was a bigger inch.

[78] So when something, then 5 inches, then 3 inches.

[79] Yeah, I think that was the 5.

[80] I don't, maybe before that was the giant plates, and I didn't get that.

[81] But it was definitely not the 3 -inch ones.

[82] Anyway, so that, you know, we got that computer.

[83] I spent the first few months just, you know, playing video games, as you would expect.

[84] I got bored of that, so I started messing around and trying to figure out how to make the thing, do other stuff.

[85] I got into exploring, programming, and a couple of years later, it got to a point where I actually wrote a game, a lot of games, and a game developer, a Japanese game developer, actually offered to buy it for me for a few hundred bucks, but for a kid in Russia.

[86] It's a big deal.

[87] It's a big deal, yeah.

[88] I did not take the deal.

[89] Wow.

[90] Integrity.

[91] I instead...

[92] Distipidity.

[93] Yes, that was not the most acute financial move that I made my life.

[94] You know, looking back at it now.

[95] I instead put it, well, you know, I had a reason.

[96] I put it online.

[97] It was, what did you call it back in the days?

[98] It was a freeware thing, right?

[99] It was not open source, but you could upload the binaries.

[100] You would put the game online.

[101] And the idea was that, you know, people like it.

[102] And then they, you know, contribute and they send you a little donations, right?

[103] So I did my quick math of, like, you know, Of course, you know, thousands and millions of people are going to play my game, send me a couple of bucks a piece, you know, should definitely do that.

[104] As I said, not the best financial.

[105] You're already playing business models at that age.

[106] Remember what language it was?

[107] What programming?

[108] It was a base.

[109] Pascal.

[110] Pascal.

[111] And they had a graphical component, so it's not text -based.

[112] Yeah, yeah, it was like, I think, you know, 320 by 200, whatever it was.

[113] I think that kind of the earlier version.

[114] VGA resolution, right?

[115] And I actually think the reason, the reason.

[116] I actually think the reason, and why this company wanted to buy it.

[117] It's not like the fancy graphics or the implementation.

[118] It was maybe the idea of actual game.

[119] The idea of the game.

[120] Well, one of the things, it's so funny, I used to play this game called Golden Axe, and the simplicity of the graphics and something about the simplicity of the music, like, it still haunts me. I don't know if that's a childhood thing.

[121] I don't know if that's the same thing for Call of Duty these days for young kids, but I still think that, that the simple, one of the games are simple, that simple purity makes for, like, allows your imagination to take over and thereby creating a more magical experience.

[122] Like now with better and better graphics, it feels like your imagination doesn't get to create worlds, which is kind of interesting.

[123] It could be just an old man on a porch, like waving at kids these days that have no respect but i still think that graphics almost get in the way of the experience i don't know flip a bird yeah i don't know i don't know if the imagination case closed i don't yeah but that that's more about games that up like that's more like tetris world where they optimally masterfully like create a fun short -term dopamine experience versus i'm more referring to like robering playing games where there's like a story you can live in it for months or years um like uh there's an elder scroll series which is probably my favorite set of games that was a magical experience and that the graphics are terrible the characters were all randomly generated but they're i don't know that's it pulls you in there's a story it's like an interactive version of an elder scroll's Tolkien world and you get to live in it i don't know i miss it It's one of the things that suck about being an adult is there's no, you have to live in the real world as opposed to the elder scrolls world.

[124] You know, whatever brings you joy, right?

[125] Minecraft is a great example.

[126] You create, like it's not the fancy graphics, but it's the creation of your own worlds.

[127] Yeah, that one is crazy.

[128] You know, one of the pitches for being a parent that people tell me is that you can like use the excuse of parenting to go back into the video game world.

[129] and like that's like you know father son father daughter time but really you just get to play video games with your kids so anyway at that time did you have any ridiculous ambitious dreams of where as a creator you might go as an engineer did you what did you think of yourself as as an engineer as a tinker or did you want to be like an astronaut or something like that you know I'm tempted to make something up about, you know, robots, engineering or, you know, mysteries of the universe, but that's not the actual memory that pops into my mind when you ask me about childhood dreams.

[130] I'll actually share the real thing.

[131] When I was maybe four or five years old, I, you know, as we all do, I thought about no one I wanted to do when I grow up.

[132] And I had this dream of being a traffic control couple.

[133] they don't have those todays I think but you know back in the 80s and you know in Russia you probably are familiar with that Lex they had these you know police officers that would stand in the middle of intersection all day and they would have their like striped back black and white batons that they would use to you know control the flow of traffic and you know for whatever reason I was strangely infatuated with this whole process and like that that was my dream that's what I wanted to do when I grew up and you know my parents both physics profs, by the way.

[134] I think we're a little concerned with that level of ambition coming from their child at that age.

[135] Well, it's an interesting.

[136] I don't know if you can relate, but I very much love that idea.

[137] I have a OCD nature that I think lends itself very close to the engineering mindset, which is you want to kind of optimize, you know, solve a problem by creating an automated solution, like a set of rules, the set of rules you can follow, and then thereby make it ultra -efficient.

[138] I don't know if that's, it was of that nature.

[139] I certainly have that.

[140] There's like SimCity and factory building games, all those kinds of things, kind of speak to that engineering mindset.

[141] Or did you just like the uniform?

[142] I think it was more of the latter.

[143] I think it was the uniform and the, you know, the striped baton that made cars go in the right directions.

[144] But I guess, you know, I did end up, I guess, you know, working on the transportation industry one way or another.

[145] No uniform, no. That's right.

[146] Maybe it was my, you know, deep inner infatuation with the, you know, traffic control batons that led to this career.

[147] Okay, what, when did you, when was the leap from programming to robotics?

[148] That happened later.

[149] That was after grad school.

[150] After, and actually, they know, cell driving cars was, I think, my first real hand.

[151] hands -on introduction to robotics.

[152] But I never really had that much hands -on experience.

[153] In school and training, I worked on applied math and physics.

[154] Then in college, I did more abstract computer science.

[155] And it was after grad school that I really got involved in robotics, which was actually selling cars.

[156] And that was a big bit flip.

[157] What grad school?

[158] So I went to grad school in Michigan, and then I did a postdoc at Stanford, which was the postdoc where I got to play with selling driving cars.

[159] Yeah, so we'll return there, but let's go back to Moscow.

[160] So, you know, for episode 100, I talked to my dad, and also I grew up with my dad, I guess.

[161] So I had to put up with them for many years.

[162] And he went to the Fistiek, or M -IPT.

[163] It's weird to say in English, because I've heard all of this in Russian.

[164] Moscow Institute of Physics and Technology and to me that was like I met some super interesting as a child I met some super interesting characters it felt to me like the greatest university in the world the most elite university in the world and just the people that I met that came out of there were like not only brilliant but also special humans it seems like that place really tested the soul, both like in terms of technically and like spiritually.

[165] So that could be just the romanticization of that place.

[166] I'm not sure.

[167] So maybe you can speak to it.

[168] But is it correct to say that you spend some time at Fistia?

[169] Yeah, that's right.

[170] Six years.

[171] I got my bachelor's and masters and physics and math there.

[172] And it's actually interesting because my dad, actually Both my parents went there, and I think all the stories that I heard, just like you, Alex, growing up about the place and, you know, how interesting and special and, you know, magical it was.

[173] I think that was a significant, maybe the main reason I wanted to go there for college, enough so that I actually went back to Russia from the U .S. I graduated high school in the U .S. You went back there.

[174] I went back there, yeah.

[175] Wow.

[176] Exactly the reaction most of my peers in college had.

[177] but, you know, perhaps a little bit stronger that, like, you know, point me out as this crazy kid.

[178] Were your parents supportive of that?

[179] My name was your previous question.

[180] They, uh, they supported me and, you know, letting me, kind of pursue my passions and the, you know, things that I was understood.

[181] That's a bold move.

[182] What was it like there?

[183] It was interesting.

[184] You know, definitely fairly hardcore on the fundamentals of, you know, math and physics and, you know, lots of good memories from, you know, from those times.

[185] So, okay.

[186] So, Stanford.

[187] How'd you get into autonomous vehicles?

[188] I had the great fortune and great honor to join Stanford's DARPA Urban Challenge team in 2006.

[189] This was a third in the sequence of the DARPA challenges.

[190] There were two grand challenges prior to that.

[191] And then in 2007, they held the DARPA Urban Challenge.

[192] So, you know, I was doing my postdoc, I had joined the team and worked on motion.

[193] planning for that competition so okay so for people who might not know I know from from a certain autonomous vehicles is a funny world in a certain circle of people everybody knows everything and in a certain circle nobody knows anything in terms of general public so it's interesting it's it's a good question of what to talk about but I do think that the urban challenge is worth revisiting it's a fun little challenge one that in But first, it, like, sparked so much, so many incredible minds to focus on one of the hardest problems of our time in artificial intelligence.

[194] So that's a success from a perspective of a single little challenge.

[195] But can you talk about, like, what did the challenge involve?

[196] So were there pedestrians, were there other cars?

[197] What was the goal?

[198] Who was on the team?

[199] How long did it take?

[200] Any fun, fun sort of specs?

[201] Sure, sure.

[202] So the way the challenge was constructed and just a little bit of backgrounding.

[203] As I mentioned, this was the third competition in that series.

[204] The first two were the grand challenge called the grand challenge.

[205] The goal there was to just drive in a completely static environment.

[206] You know, you had to drive in a desert.

[207] That was very successful.

[208] So then DARPA followed with what they called the urban challenge where the goal was to build vehicles that could operate in more dynamic environments and share them with other vehicles.

[209] There were no pedestrians there, but what DARPA did is they took over an abandoned Air Force base, and it was kind of like a little fake city that they built out there.

[210] And they had a bunch of robots, cars that were autonomous in there, all at the same time, mixed in with other vehicles driven by professional drivers.

[211] And each car had a mission.

[212] And so there's a crude map that they received the beginning and they had a mission and go here and then there and over here.

[213] And they kind of all were sharing this environment at the same time they had to interact with each other.

[214] They had to interact with the human drivers.

[215] So it's this very first, very rudimentary version of a self -driving car that could operate in an environment shared with other dynamic actors that, as you said, you know, really, in many ways, you know, kick started this whole industry.

[216] Okay, so who was on the team, and how'd you do?

[217] I forget.

[218] It came in second.

[219] Perhaps that was my contribution to the team.

[220] I think the Stanford team came in first in the DARPA challenge, but then I joined the team.

[221] You were the one with the bug and the code.

[222] I mean, do you have sort of memories of some particularly challenging things or, you know, one of the cool things, it's not, you know, this isn't a product, this isn't the thing that, you know, it's, you know, it's.

[223] You have a little bit more freedom to experiment so you can take risks, and so you can make mistakes.

[224] Is there interesting mistakes?

[225] Is there interesting challenges that stand out to you?

[226] Some, like, taught you a good technical lesson or a good philosophical lesson from that time?

[227] Yeah, definitely, definitely a very memorable time.

[228] Not really a challenge, but, like, one of the most vivid memories that I have from the time.

[229] And I think that was actually one of the days that really got me hooked on this whole field was the first time I got to run my software on the car.

[230] And I was working on a part of our planning algorithm that had to navigate in parking lots.

[231] So it was something that called free space motion planning.

[232] So the very first version of that was we tried on the car.

[233] It was on Stanford's campus in the middle of the night.

[234] night, and you know, had this little, you know, course constructed with cones in the middle of a parking lot, so we were there in like 3 a .m. You know, by the time we got the code to, you know, compile and turnover.

[235] And, you know, it drove.

[236] I could actually do something quite reasonable.

[237] And, you know, it was, of course, very buggy at the time and had all kinds of problems.

[238] But it was pretty darn magical.

[239] I remember going back and, you know, you know, later at night trying to fall asleep and just, you know, being on.

[240] able to fall asleep for the rest of the night, just my mind was blown.

[241] And that's what I've been doing ever since for more than a decade.

[242] In terms of challenges and, you know, it's interesting memories.

[243] Like on the day of the competition, it was pretty nerve -wrecking.

[244] I remember standing there with Mike Montemarillo, who was the software lead and wrote most of the code.

[245] I think I did one little part of the planner.

[246] Mike, you know, incredibly did pretty much the rest of it with, you know, a bunch of other incredible people.

[247] But I remember standing on the day of the competition, you know, watching the car with Mike and, you know, cars are completely empty, right?

[248] They're all there lined up in the beginning of the race.

[249] And then, you know, DARPA sends them, you know, on their mission, one by one.

[250] So then leave.

[251] And like, you just, they had these sirens, right?

[252] They all had their different silence, silence, right?

[253] Each siren had its own personality, if you will.

[254] So, you know, off the go.

[255] And you don't see them.

[256] And then every once in a while, they, you know, come a little bit closer to where the audiences and you can kind of hear the sound of your car and it seems to be moving along so that gives you hope and then it goes away and you can't hear it for too long you start getting anxious right so it's a little bit like you know sending your kids to college and like you know kind of you invested in them you hope you you you know you build it properly but like it's still anxiety inducing so that was an incredibly fun a few days in terms of you know bugs as you mentioned you know one that that was my bug that caused the loss of the first place.

[257] There's still a debate that, you know, occasionally have with people on the CMU team.

[258] CMU came first, I should mention.

[259] That - C -U haven't heard of them, but yeah.

[260] It's some, you know, little school.

[261] It's a small school somewhere.

[262] It's, you know, really a glitch that, you know, they happen to succeed at something robotics -related.

[263] Very scenic, though.

[264] So most people go there for the scenery.

[265] Yeah, that's a beautiful campus.

[266] I apologize.

[267] Unlike Stanford.

[268] So for people, yeah, that's true, unlike Stanford, for people who don't know CMU is, one of the great robotics and sort of artificial intelligence universities in the world.

[269] CMU, Carnegie -Bellon University.

[270] Okay, sorry, go ahead.

[271] Good PSA.

[272] So in the part that I contributed to, which was navigating parking lots, and the way that part of the mission worked is in a parking lot, you would get from DARPA an outline of the map.

[273] You basically get this giant polygon that defined the perimeter of the parking lot.

[274] And there would be an entrance and maybe multiple interests or access to it.

[275] And then you would get a goal within that open space, X, Y, heading, where the car had to park.

[276] It had no information about the optical, so obstacles that the car might encounter there.

[277] So it had to navigate completely free space from the entrance to the parking lot into that parking space.

[278] And then once parked there, it had to exit the parking lot.

[279] while, of course, in counting and reasoning about all the obstacles that it encounters in real time.

[280] So our interpretation, or at least my interpretation of the rules, was that you had to reverse out of the parking spot, and that's what our cars did, even if there's no obstacle in front.

[281] That's not what CMU's car did, and it just kind of drove right through.

[282] So there's still a debate, and of course, you know, if you stop and then reverse out and go out the different way, that cost you some time.

[283] So there's still a debate whether, you know, it was my point.

[284] implementation that cost us extra time, or whether it was, you know, CMU violating an important rule of the competition.

[285] And, you know, I have my own opinion here.

[286] In terms of other bugs, and like I have to apologize to Mike Montemarillo for sharing this on air, but it is actually one of the more memorable ones.

[287] And it's something that's kind of become a bit of a metaphor and a label in the industry since then, I think, at least in some circles, it's called a victory circle or victory lap.

[288] And our cars did that.

[289] So in one of the missions in the urban challenge, in one of the courses, there was this big oval right by the start and finish of the race.

[290] So the ARPA had a lot of the missions would finish in that same location.

[291] And it was pretty cool because you could see the cars come by, you know, kind of finish that part leg of the trip, that leg of the mission, and then, you know, go on and finish the rest of it.

[292] And other vehicles, you know, the other vehicles, would, you know, come, hit their waypoint and, you know, exit the oval and off they would go.

[293] Our car, on the hand, which hit the checkpoint, and then it would do an extra lap around the oval and only then, you know, leave and go in its merry way.

[294] So over the course of, you know, the full day, it accumulated some extra time.

[295] And the problem was that we had a bug where it wouldn't, you know, start reasoning about the next waypoint and plan a route to get to that next point until it hit a previous one.

[296] And in that particular case, by the time you hit that one, it was too late for us to consider the next one and kind of chameka lane change so every time it would do like an extra lap so and that's the the Stanford victory lap the victory lap oh that's there's I feel like there's something philosophically profound in there somehow but uh I mean ultimately everybody is a winner in that kind of competition and it it's led to sort of famously to the creation of uh Google self -driving car project and now Waymo so can we uh give an over of how is way more born, how is the Google self -driving car project born?

[297] What is the mission?

[298] What is the hope?

[299] What is it is the engineering kind of set of milestones that it seeks to accomplish?

[300] There's a lot of questions in there.

[301] Yeah.

[302] I don't know.

[303] But you're right.

[304] Kind of the DARPA Urban Challenge and the previous DARPA Grand Challenges can lead, I think, to a very large degree to that next step.

[305] And then Larry and Sergey, Larry Page and Sergey Brin, Google Hunter scores, saw that competition and believed in the technology.

[306] So the Google self -driving car project was born.

[307] At that time, and we started in 2009, it was a pretty small group of us, about a dozen people who came together to work on this project at Google.

[308] At that time, we saw that incredible early result in the DARPA urban challenge.

[309] I think we're all incredibly excited about where we got to, and we believed in the future of the technology, but we still had a very rudimentary understanding of the problem space.

[310] So the first goal of this project in 2009 was to really better understand what we're up against.

[311] And with that goal in mind, when we started the project, we created a few milestones for ourselves that maximized learnings, if you will.

[312] The two milestones were, you know, one was to drive 100 ,000 miles in autonomous mode, which was at that time, you know, orders of magnitude that more than anybody has ever done.

[313] And the second milestone was to drive 10 routes.

[314] Each one was 100 miles long.

[315] There were specifically chosen to be kind of extra spicy, you know, extra complicated and sample the full complexity of that domain.

[316] And you had to drive each one from beginning to end with no intervention, no human intervention.

[317] So you would get to the beginning of the course.

[318] You would press the button, that would engage in autonomy, and you had to go for 100 miles beginning to end with no interventions.

[319] And it sampled, again, the full complexity of driving conditions.

[320] Some were on freeways.

[321] We had one route that went all through all, freeways and all the bridges in the Bay Area.

[322] We had some that went around Lake Tahoe and kind of mountains roads.

[323] We had some that drove through dense urban environments like in downtown Palo Alto and through San Francisco.

[324] So it was incredibly interesting to work on.

[325] And it took us just under two years, about a year and a half a little bit more to finish both of these milestones.

[326] And in that process, A, it was an incredible amount of fun, probably the most fun I had in my professional career, and you're just learning so much.

[327] You are, you know, the goal here is to learn a prototype.

[328] You're not yet starting to build a production system, right?

[329] So you just, you were, you know, this is when you're kind of, you know, working 24 -7 and, you're hacking things together.

[330] And you also don't know how hard this is.

[331] I mean, that's the point.

[332] Like, so, I mean, that's an ambitious, if I put myself in that mindset, even still, that's a, a really ambitious set of goals like just those two picking it's picking 10 different difficult spicy challenges and then having zero interventions so like not saying gradually we're going to like you know over a period of 10 years we're going to have a bunch of roots and gradually reduce the number of interventions you know that literally says like by as soon as possible want to have zero and on hard roads.

[333] So to me, if I was facing that, it's unclear whether that takes two years or whether that takes 20 years.

[334] It took us under two.

[335] I guess that speaks to a really big difference between doing something once and having a prototype where you're going after learning about the problem versus how you go about engineering a product that, where you look at, you know, do you properly do evaluation, you look at metrics, you know, drive down, and you're confident that you can do that at a hundred.

[336] And I guess that's the, you know, why it took a dozen people, you know, 16 months or a little bit more than that, back in 2009 and 2010, with the technology of, you know, the more than a decade ago, that amount of time to achieve that milestone of 10 routes, 100 miles each and no interventions, and, you know, it took us a little bit longer to get to, you know, a full driverless product that customers used.

[337] That's another really important moment.

[338] Is there some memories of technical lessons or just one, like, what did you learn about the problem of driving from that experience?

[339] I mean, we can now talk about, like, what you learned from modern -day Waymo, but I feel like you may have learned some profound things in those early days, even more so.

[340] Because it feels like what Waymo is now is to trying to, you know, how to do scale, how to make sure you create a product, how to make sure it's like safety and all those things, which is all fascinating challenges.

[341] But like you were facing the more fundamental philosophical problem of driving in those early days.

[342] Like, what the hell is driving as an autonomous?

[343] Or maybe I'm again romanticizing it, but.

[344] Is there some valuable lessons you picked up over there at those two years?

[345] A ton.

[346] The most important one is probably that we believe that it's doable.

[347] And we've gotten far enough into the problem that, you know, we had, I think, only a glimpse of the true complexity of that domain.

[348] It's a little bit like, you know, climbing a mountain where you kind of see the next peak and you think that's kind of the summit but then you get to that and you kind of see that this is just the start of the journey but we've tried we've sampled enough of the problem space and we've made enough rapid success even you know with technology of 2009 to 2010 that it gave us confidence to then you know pursue this as a real product so okay so the next step.

[349] You mentioned the milestones that you had in those two years.

[350] What are the next milestones that then led to the creation of Waymo and beyond?

[351] It was a really interesting journey.

[352] And Waymo came a little bit later.

[353] Then we completed those milestones in 2010.

[354] That was the pivot when we decided to focus on actually building a product using this technology.

[355] The initial couple years after that we were focused on a freeway, what you would call a driver assist, maybe on an L3 driver assist program.

[356] Then around 2013, we've learned enough about the space and have thought more deeply about, you know, the product that we wanted to build that we pivoted.

[357] We pivoted towards this vision of building a driver and deploying it fully driverless vehicles without a person and that that's the path that we've been on since then and very, it was exactly the right decision for us.

[358] So there was a moment where you're also considered, like, what is the right trajectory here?

[359] What is the right role of automation in the task of driving?

[360] There was still, it wasn't from the early days, obviously you want to go fully autonomous.

[361] From the early days, it was not.

[362] I think it was around 2013, maybe that we've, that became very clear, and we made that pivot, and it also became very clear.

[363] and that it's even the way you go building a driver system is fundamentally different from how you go building a fully driverless vehicle.

[364] So, you know, we've pivoted towards the ladder, and that's what we've been working on ever since.

[365] And so that was around 2013.

[366] Then there's a sequence of really meaningful for us, really important defining milestones since then.

[367] 15, we had our first, actually the world's first, fully driverless ride on public roads.

[368] It was in a custom -built vehicle that we had.

[369] I must have seen those.

[370] We called them the Firefly, that funny -looking, marshmallow -looking thing.

[371] And we put a passenger, his name was Steve Mann, a great friend of our project from the early days.

[372] the man happens to be blind.

[373] So we put him in that vehicle.

[374] The car had no steering wheel, no pedals.

[375] It was an uncontrolled environment.

[376] No lead or chase cars, no police escorts.

[377] And we did that trip a few times in Austin, Texas.

[378] So that was a really big milestone.

[379] But that was in Austin.

[380] Yeah.

[381] Cool.

[382] Okay.

[383] And we only, but at that time, it took a tremendous amount of engineering.

[384] It took a tremendous amount of validation to get to that point.

[385] but, you know, we only did it a few times.

[386] I only did that.

[387] It was a fixed route.

[388] It was not kind of a controlled environment, but it was a fixed route, and we only did it a few times.

[389] Then in 2016, end of 2016, beginning of 2017, is when we founded Waymo, the company.

[390] That's when we kind of, that was the next phase of the project where we believed in kind of the commercial vision of this technology.

[391] And it made sense to create an independent entity within that alphabet umbrella to pursue this product at scale.

[392] Beyond that, in 2017, later in 2017, it was another really huge step for us, really big milestone where we started, I think it was October of 2017, where when we started regular driverless operations on public roads, that first day of operations.

[393] we drove in one day, in that first day, 100 miles in, you know, driverless fashion.

[394] And then we've, the most important thing about that milestone was not that, you know, 100 miles in one day, but that it was the start of kind of regular ongoing driverless operations.

[395] And when you say driverless, it means no driver.

[396] That's exactly right.

[397] So on that first day, we actually had a mix.

[398] And in some, we didn't want to, like, you know, be on YouTube and Twitter that same day.

[399] So in many of the rides.

[400] We had somebody in the driver's seat, but they could not disengage.

[401] Like the car, not disengaged.

[402] But actually, on that first day, some of the miles were driven and just completely empty driver's seat.

[403] And this is the key distinction that I think people don't realize that oftentimes when you talk about autonomous vehicles, there's often a driver in the seat that's ready to take over, what's called a safety driver.

[404] And then Waymo is really one of the only companies, at least that I'm aware of, or at least as like boldly and carefully and all of that, is actually has cases, and now we'll talk about more and more where there's literally no driver.

[405] So that's another, the interesting case of where the driver's not supposed to disengage, that's like a nice middle ground.

[406] They're still there, but they're not supposed to disengage.

[407] but really there's the case when there's no, okay, there's something magical about there being nobody in the driver's seat.

[408] Like, just like to me, you mentioned the first time you wrote some code for free space navigation of the parking lot, that was like a magical moment.

[409] To me, just sort of as an observer of robots, the first magical moment is seeing an autonomous vehicle turn like make a left turn like apply sufficient torque to the steering wheel to where like there's a lot of rotation and for some reason and there's nobody in the driver's seat for some reason that that communicates that here's a being with power that makes a decision there's something about like the steering wheel because we perhaps romanticize the notion of the steering wheel it's so essential to our conception our 20th century conception of a car and it turning the steering wheel with nobody in driver's seat that to me I think maybe to others it's really powerful like this thing is in control and then there's this leap of trust that you give like I'm going to put my life in the hands of this thing that's in control so in that sense when there's no driver in the driver seat that's a magical moment for robots so I'm I gotten a chance to last year to take a ride in a Waymo vehicle.

[410] And that was the magical moment.

[411] There's like nobody in the driver's seat.

[412] It's like the little details.

[413] You would think it doesn't matter whether there's a driver or not.

[414] But like if there's no driver and the steering wheel is turning on its own, I don't know.

[415] That's magical.

[416] It's absolutely magical.

[417] I've taken many of these rides and a completely empty car.

[418] No human in the car pulls up.

[419] You know, you call it on your cell phone, it pulls up.

[420] You get in, it takes you on its way.

[421] There's nobody in the car but you, right?

[422] That's something called, you know, fully driverless, you know, our writer -only mode of operation.

[423] Yeah, it is magical.

[424] It is, you know, transformative.

[425] This is what we hear from our riders.

[426] It kind of really changes your experience.

[427] And not like that that really is what unlocks the real potential of this technology.

[428] But, you know, coming back to our.

[429] journey, you know, that was 2017 when we started truly driverless operations.

[430] Then in 2018, we've launched our public commercial service that we called Waymo One in Phoenix.

[431] In 2019, we started offering truly driverless rider -only rights to our early rider population of users.

[432] And then, you know, 2020 has also been a pretty interesting year, one of the first months, less about technology, but more about the maturing and the growth of Waymo as a company.

[433] We raised our first round of external financing this year.

[434] We're part of Alphabet, so obviously we have access to significant resources, but as kind of on the journey of Waymo maturing as a company, it made sense for us to, you know, partially go externally in this round.

[435] So, you know, we're raised.

[436] about $3 .2 billion worth from, you know, that round.

[437] We've also, you know, started putting our fifth generation of our driver, our hardware that is on the new vehicle, but it's also a qualitatively different set of self -driving hardware that is now on the JLR pace.

[438] So that was a very important step for us.

[439] Hardware specs, fifth generation, I think it would be fun.

[440] Maybe I apologize if I'm interrupting, but maybe talk about maybe the generations with a focus on what we're talking about in the fifth generation in terms of hardware specs.

[441] Like what's on this car?

[442] Sure.

[443] So we separated out, you know, the actual car that we are driving from the self -driving hardware we put on it.

[444] Right now we have, so this is, as I mentioned, the fifth generation.

[445] You know, we've gone through, we started, you know, building our own hardware.

[446] Many, many years ago, and that Firefly vehicle also had the hardware suite that was mostly design, engineered and built in -house.

[447] Liders are one of the more important components that we design and build from the ground up.

[448] So on the fifth generation of our drivers of our self -driving hardware that we're switching to right now, we have as with previous generations in terms of sensing we have lighters cameras and radars and we have a pretty beefy computer that processes all that information and makes decisions in real time on board the car so in all of the and it's really a qualitative jump forward in terms of the capabilities and the various parameters and the specs of the hardware compared to what we had before and compared to what you can kind of of get off the shelf in the market today.

[449] Meaning from fifth to fourth or from fifth to first?

[450] Definitely from first to fifth, but also from the world's dumbest question.

[451] Definitely from fourth to fifth.

[452] As well as this, that last step is a big step forward.

[453] So everything's in -house, so like LiDAR is built in house and cameras are built in house?

[454] You know, it's different.

[455] You know, we work with partners.

[456] There's some components that, you know, We, you know, get from our manufacturing and, you know, supply chain partners.

[457] What exactly is in -house is a bit different.

[458] We do a lot of, you know, custom design on all of our sensing models, lighters, radars, cameras.

[459] You know, exactly there's, lighters are almost, you know, exclusively in -house.

[460] And some of the technologies that we have, some of the fundamental technologies there are completely unique to Waymo.

[461] that is also largely true about radars and cameras it's a little bit more of a mix in terms of what we do ourselves versus what we get from partners is there something super sexy about the computer that you can mention that's not top secret like uh for people who enjoy computers for i mean uh see there's there's a lot of machine learning involved but there's a lot of just basic compute there's you have to uh probably do a lot of signal processing on all the sensors, you have to integrate everything has to be in real time.

[462] There's probably some kind of redundancy type of situation.

[463] Is there something interesting you can say about the computer for the people who love hardware?

[464] It does have all of the characteristics, all the properties that you just mentioned.

[465] Redundancy, very beefy compute for general processing as well as, you know, inference and ML models.

[466] It is some of the more sensitive stuff that, you know, I don't want to get into for IP reasons, but we've shared a little bit in terms of the specs of the sensors that we have on the car.

[467] You know, we actually shared some videos of what our lighter sees, lighters see in the world.

[468] We have 29 cameras.

[469] We have five lighters.

[470] We have six radars on these vehicles.

[471] And you can kind of get a feel for the amount of data that they're producing.

[472] That all has to be processed in real time to do perception, to do complex.

[473] reasoning.

[474] It kind of gives you some idea of how beefy those computers are, but I don't want to get into specifics of exactly how we build up.

[475] Okay, well, let me try some more questions that you can get into the specifics of like GPU -wise.

[476] Is that something you can get into?

[477] You know, I know that Google works with DPUs and so on.

[478] I mean, for machine learning folks, it's kind of interesting.

[479] Or is there no, how do I ask it?

[480] I've been talking to people in the government about UFOs and they don't want to answer any questions.

[481] So this is how I feel right now asking about GPS.

[482] But is there something interesting that you could reveal or is it just, you know, or leave it up to our imagination some of the compute?

[483] Is there any, I guess, is there any fun trickery?

[484] Like I talked to Chris Latner for a second time and he has a key person about TPUs and there's a lot of fun stuff going on in Google in terms of hardware that optimizes for machine learning.

[485] Is there something you can reveal?

[486] in terms of how much, you mentioned customization, how much customization there is for hardware for machine learning purposes.

[487] I'm going to be like that government, you know, you've got a person, body of foes.

[488] But I guess I will say that it's really, compute is really important.

[489] We have very data -hungry and compute -hungry ML models all over our stack, and this is where, you know, both being part of Alphabet as well as designing our own sensors and the entire hardware suite together, where on one hand you get access to really rich raw sensor data that you can pipe from your sensors into your compute platform and build the whole pipe from sensor raw sensor data to the big compute, as then have the massive compute to process all that data.

[490] This is where we're finding that having a lot of control.

[491] of that hardware part of the stack is really advantageous.

[492] One of the fascinating magical places to me, again, might not be able to speak to the details, but it is the other compute, which is like, you know, we're just talking about a single car, but the, you know, the driving experience is a source of a lot of fascinating data, and you have a huge amount of data coming in on the car, and, you know, the infrastructure of storing some of that data to then train or to analyze or so on.

[493] That's a fascinating piece of it that I understand a single car.

[494] I don't understand how you pull it all together in a nice way.

[495] Is that something that you can speak to in terms of the challenges of seeing the network of cars and then bringing the data back and analyzing things like edge cases of driving, be able to learn on them to improve the system, to see where things went wrong, where things went right, and analyze all that kind of stuff.

[496] Is there something interesting there from an engineering perspective?

[497] Oh, there's an incredible amount of really interesting work that's happening there, both in the real -time operation of the fleet of cars and the information that they exchanged with each other in real -time to make better decisions, as well as on the kind of the off -board component, where you have to deal with massive amounts of data for training your ML models, evaluating the ML models, for simulating the entire system, and for evaluating your entire system.

[498] This is where being part of Alphabet has been tremendously advantageous.

[499] We consume an incredible amount of compute for ML infrastructure.

[500] We build a lot of custom frameworks to get good at data mining, finding the interesting age cases for training and for evaluation of the system for both training and evaluating some components and your sub parts of the system on various ML models as well as evaluating the entire system and simulation.

[501] Okay, is that first piece that you mentioned that cars communicating to each other essentially, I mean, through perhaps through a centralized point.

[502] But what, that's fascinating too.

[503] How much does that help you?

[504] Like if you imagine right now the number of way more vehicles is whatever x i don't know if you can talk to what that number but it's it's not in the hundreds of millions yet and imagine if the whole world is way more vehicles uh like that changes potentially the power of connectivity like the more cars you have i guess actually if you look at phoenix because there's enough vehicles uh there's enough when there's like some level of density you can start to probably do some really interesting stuff with the fact that cars can negotiate can be can communicate with each other and thereby make decisions is there something interesting there that you can talk to about like how does that help with the driving problem from as compared to just a single car solving the driving problem by itself uh yeah it's it's a spectrum i at first i say that yeah it's it helps uh and it helps in various ways, but it's not required right now.

[505] The way we build our system, like each cars can operate independently.

[506] They can operate with no connectivity.

[507] So I think it is important that you have a fully autonomous, fully capable driver that computerized driver that each car has.

[508] Then, you know, they do share information and they share information in real time.

[509] It really helps.

[510] So the way we do this today is, you know, whenever one car encounter something interesting in the world, whether it might be an accident or a new construction zone, that information immediately gets uploaded over the air and is propagated to the rest of the fleet.

[511] And that's kind of how we think about maps as priors in terms of the knowledge of our fleet of drivers that is distributed across the fleet.

[512] And it's updated in real time.

[513] So that's one use case.

[514] You can imagine as the density of these vehicles go up, that they can exchange more information in terms of what they're planning to do and start influencing how they interact with each other, as well as potentially sharing some observations to help with, if you have enough density of these vehicles where one car might be seeing something that another is relevant to another car that is very dynamic.

[515] It's not part of you're updating your static prior off the map of the world, but it's more of a dynamic information that could be relevant to the decisions that another car is making real -time, so you can see them exchanging that information, and you can build on that.

[516] But again, I see that as an advantage, but it's not a requirement.

[517] So what about the human in the loop?

[518] So when I got a chance to drive with a ride in a Waymo, you know, there's customer service.

[519] So like there's somebody that's able to dynamically, like, tune in and help you out, what role does the human play in that picture?

[520] That's a fascinating.

[521] Like, you know, the idea of teleoperation be able to remotely control a vehicle.

[522] So here what we're talking about is, like, like, frictionless, like a human being able to, in a frictionless way, sort of help you out.

[523] I don't know if they're able to actually you control the vehicle?

[524] Is that something you can talk to?

[525] Yes.

[526] To be clear, we don't do teleoperation.

[527] I'm going to believe in teleoperation for a reason is that's not what we have on our cars.

[528] We do, as you mentioned, have a version of customer support.

[529] We call it Live Health.

[530] In fact, we find it that it's very important for our writer experience, especially if it's your first trip.

[531] You've never been in a fully driverless, right, or only way more vehicle.

[532] You get in.

[533] There's nobody there.

[534] So you can imagine having all kinds of questions in your head, like how this thing works.

[535] So we've put a lot of thought into kind of guiding our riders, our customers through that experience, especially for the first time.

[536] They get some information on the phone if the fully driverless vehicle is used to service their trip.

[537] When you get into the car, we have an in -car screen and audio that kind of guides them and explains what to expect.

[538] they also have a button that they can push that will connect them to, you know, a real life human being that they can talk to, right, about this whole process.

[539] So that's one aspect of it.

[540] There is, you know, I should mention that there is another function that humans provide to our cars, but it's not teleoperation.

[541] You can think of it a little bit more like, you know, fleet assistance, kind of like, you know, traffic control that you have, where our cars, again, they're responsive.

[542] on their own for making all of the decisions, all of the driving decisions that don't require connectivity.

[543] Anything that is safety or latency critical is done purely autonomously by on -board system.

[544] But there are situations where, you know, if connectivity is available, in a car encounters a particularly challenging situation, you can imagine like a super hairy scene of an accident.

[545] The cars will do their best, they will recognize that it's an off -nominal situation.

[546] they will do their best to come up with the right interpretation, the best course of action in that scenario.

[547] But if the connectivity is available, they can ask for confirmation from, you know, a human assistanter to kind of confirm those actions and perhaps provide a little bit of contextual information and guidance.

[548] So October 8th was when you're talking about, was Waymo launched the fully self, the public version of its fully driverless, that's the right term, I think, service in Phoenix.

[549] Is that October 8th?

[550] That's right.

[551] It was the introduction of fully driverless rider -only vehicles into our public Waymo One service.

[552] Okay, so that's amazing.

[553] So it's like anybody can get into Waymo in Phoenix?

[554] That's right.

[555] So we previously had early people in our early rider program taking fully driverless rides in Phoenix.

[556] And just this, a little while ago, we opened on October 8th, we opened that mode of operation to the public.

[557] So I can download the app and go on the right.

[558] There is a lot more demand right now for that service.

[559] And then we have capacity.

[560] So we're kind of managing that, but that's exactly the way you described it.

[561] Yeah, well, that's interesting.

[562] So there's more demand than you can handle.

[563] Like what has been reception so far?

[564] like what i mean okay so you know that's this is a product right that's a whole nother discussion of like how compelling of a product it is great but it's also like one of the most kind of transformational technologies of the 21st century so there it's also like a tourist church like it's fun to you know to be a part of it so it'd be interesting to see like what do people say what do people what have been the feedback so far you know still early days But so far, the feedback has been incredible, incredibly positive.

[565] We asked them for feedback during the ride.

[566] We asked them for feedback after the ride as part of their trip.

[567] We asked them some questions.

[568] We asked them to rate the performance of our driver.

[569] Most by far, most of our drivers give us five stars in our app, which is absolutely great to see.

[570] And they're also giving us feedback on things we can improve.

[571] And that's one of the main reasons we're doing this.

[572] Phoenix and over the last couple of years and every day today we are just learning a tremendous amount of new stuff from our users there's no substitute for actually doing the real thing actually having a fully driverless product out there in the field with you know users that are actually going to paying us money to get from point A to point B so this is a legitimate like there's a paid service that's right and the idea is you use the app to go from point A to point B And then what are the A's, what's the freedom of the starting and ending places?

[573] It's an area of geography where that service is enabled.

[574] It's a decent size of geography of territory.

[575] It's actually larger than size of San Francisco.

[576] And within that, you have full freedom of, you know, selecting where you want to go.

[577] Of course, there's some, and on your app, you get a map.

[578] you tell the car where you want to be picked up and where you want the car to pull over and pick you up and then you tell it where you want to be dropped off and of course there are some exclusions right you want to be you know where in terms of where the car is allowed to pull over right so that you can't do but besides that it's amazing it's not like a fixed I don't know maybe that's what's the question behind your question but it's not a preset set of yeah so within the geographic constraints within that area anywhere else it can be you can be picked up and dropped off anywhere.

[579] That's right.

[580] And people use them on like all kinds of trips.

[581] They, we have, and we have an incredible spectrum of riders.

[582] We have, I think the youngest, actually have car seats them and we have, you know, people taking their kids and rides.

[583] I think the youngest riders we had on cars are, you know, one or two years old, you know, and the full spectrum of use cases.

[584] People can take them to, you know, schools, to, you know, go grocery shopping, to restaurants, to bars, you know, run errands, you know, go shopping, et cetera, et cetera.

[585] You can go to your office, right?

[586] like the full spectrum of use cases.

[587] And people, you're going to use them in their daily lives to get around.

[588] And we see all kinds of really interesting use cases.

[589] And that is providing us incredibly valuable experience that we then used to improve our product.

[590] So as somebody who's been done a few long rants with Joe Rogan and others about the toxicity of the internet and the comments, and the negativity in the comments.

[591] I'm fascinated by feedback.

[592] I believe that most people are good and kind and intelligent and can provide, like, even in disagreement, really fascinating ideas.

[593] So on the product side, it's fascinating to me, like, how do you get the richest possible user feedback, like, to improve?

[594] What are the channels that you use to measure?

[595] Because, like, you're no longer, that's one of the magical things about autonomous vehicles is it's not like it's frictionless interaction with the human so like you don't get to you know it's just giving a ride so like how do you get feedback from people to in order to improve yeah great question various mechanisms so as part of the normal flow we ask people for feedback they as the car is driving around we have on the phone and in the car and we have a touchscreen in the car you can actually click some buttons and provide real -time feedback on how the car is doing and how the car is handling a particular situation, both positive and negative.

[596] So that's one channel.

[597] We have, as we discussed, customer support or life help, where if a customer wants to, it has a question or he has some sort of concern, they can talk to a person in real -time.

[598] So that is another mechanism that gives us feedback.

[599] At the end of a trip, we also ask them how things went.

[600] They give us comments and, you know, star rating.

[601] And, you know, if it's, we also, you know, ask them to explain what, you know, one well and, you know, what could be improved.

[602] And we, we have, our writers are providing, you know, very rich feedback there.

[603] A lot, a large fraction is very passionate and very excited about this technology.

[604] So we get really good feedback.

[605] We also run UXR studies, right?

[606] You know, specific and that are kind of more, you know, go more depth and we will run both kind of lateral and longitudinal studies where we have deeper engagement with our customers.

[607] You know, we have our user experience research team tracking over time.

[608] That's things about legitimacy.

[609] That's cool.

[610] That's exactly right.

[611] And, you know, that's another really valuable feedback, a source of feedback.

[612] And we're just covering a tremendous amount, right?

[613] People go grocery stropping and they like want to load, you know, 20 bags of groceries in our cars.

[614] And like that's one workflow that you maybe don't, you know, think about, you know, getting just right when you're building the driverless product.

[615] I have people, like, you know, who bike as part of their trip.

[616] So they, you know, bike somewhere.

[617] Then they get on our cars.

[618] They take a part of their bike that load into our vehicle.

[619] Then they go, and that's, you know, how they, you know, where we want to pull over and how that, you know, get in and get out.

[620] Process works, provides us, you know, useful feedback.

[621] In terms of what makes a good pickup and drop off location, we get really valuable feedback.

[622] In fact, we had to do some really interesting work with high -definition maps and thinking about walking directions.

[623] If you imagine you're in a store, in some giant space, and then, you know, you want to be picked up somewhere.

[624] Like, if you just drop a pin at a current location, which is maybe in the middle of a shopping mall, like, what's the best location for the car to come you up.

[625] And you can have simple heuristics where you just kind of take your, you know, you clean in distance and find the nearest spot where the car can't pull over that's closest to you.

[626] But oftentimes, that's not the most convenient one.

[627] You know, I have many anecdotes where that heuristic breaks in horrible ways.

[628] One example that, you know, I often mention is somebody wanted to be, you know, dropped off in Phoenix.

[629] And, you know, we car picked a location that was close, the closest to their, you know, where the pin was dropped on the map in terms of, you know, latitude and longitude.

[630] But it happened to be on the other side of a parking lot that had this row of cacti.

[631] And the poor person had to, like, walk all around the parking lot to get to where they wanted to be in 110 degree heat.

[632] So that, you know, that was a buck.

[633] So then, you know, we took all, take all of these, all of that feedback from our users and incorporate it into our system and, you know, improve it.

[634] Yeah, I feel like that's, like, requires.

[635] aGI to solve the problem of like when you're which is a very common case when you're in a big space of some kind like apartment building it doesn't matter it's some large space and then you call the like the waymo from there right like whatever doesn't matter right chair vehicle and like where is the pin supposed to drop i feel like that's i you don't think i think that requires a GI.

[636] I'm going to, in order to solve.

[637] Okay, the alternative, which I think the Google search engine is taught, is like there's something really valuable about the perhaps slightly dumb answer, but a really powerful one, which is like what was done in the past by others.

[638] Like, what was the choice made by others?

[639] That seems to be, like in terms of Google search, when you have like billions of searches that you could see which like when they recommend what you might possibly mean they suggest based on not some machine learning thing which they also do but like on what was successful for others in the past and finding a thing that they were happy with is that integrated at all way more like what what pickups worked for others it is i think you're exactly right so there's uh real it's an interesting problem uh naive solutions have interesting failure modes.

[640] So there's definitely lots of things that can be done to improve and both learning from what works, but doesn't work in actual, from getting richer data and getting more information about the environment and richer maps.

[641] But you're absolutely right that there's something, I think there's some properties of solutions that in terms of the effect, that they have on the users so much, much, much, much better than others, right?

[642] And predictability and understandability is important.

[643] So you can have maybe something that is not quite as optimal, but is very natural and predictable to the user and kind of works the same way all the time.

[644] And that matters.

[645] That matters a lot for the user experience.

[646] But to get to the basics, the pretty fundamental property is that the car actually arrives where you told it to arrive.

[647] Like you can always, you know, change it, see it on the map.

[648] and you can move it around if you don't like it.

[649] But that property that the car actually shows up on the pin is critical, which, you know, where compared to some of the human -driven analogs, I think, you know, you can have more unpredictability.

[650] It's actually the fact, if I have a little bit of a detour here, I think the fact that it's, you know, your phone and the cars, two computers talking to each other can lead to some really interesting things we can do in terms of the user interfaces.

[651] Both in terms of function, like the car actually shows up exactly where you told it you want it to be, but also some really interesting things in the user interface.

[652] I think as the car is driving, as you call it and it's on the way to come and pick you up, of course, you get the position of the car and the route on the map, and they actually follow that route, of course, but it can also share some really interesting information about what is doing.

[653] So, you know, our cars, as they are coming to pick you up, if a car is coming up to a stop sign, it will actually show you that, like, it's there sitting because it's at a stop sign, or a traffic light will show you that it's got, you know, sitting at a red light.

[654] So, you know, they look little things, right?

[655] But it's, I find those little touch touches really interesting, really magical.

[656] And it's just, you know, little things like that that you can do to kind of delight your users.

[657] You know, this makes me think of, um, there's some products that I just love.

[658] Like, there's, There's a company called Rev. Rev .com where, like for this podcast, for example, I can drag and drop a video, and then they do all the captioning.

[659] It's humans doing the captioning.

[660] But they automate everything of connecting you to the humans, and they do the captioning and transcription.

[661] It's all effortless.

[662] And I remember when I first started using them, it was like, life is good.

[663] Like, because it was so painful to figure that out earlier.

[664] The same thing with something called Isotope RX.

[665] This company I used for cleaning up audio, like the sound cleanup they do.

[666] It's like drag and drop and it just cleans everything up very nicely.

[667] Another experience like that I had with Amazon One Click purchase first time.

[668] I mean, other places do that now, but just the effortlessness of purchasing, making it frictionless.

[669] it kind of communicates to me like I'm a fan of design I'm a fan of products that you can just create a really pleasant experience that the simplicity of it the elegance just makes you fall in love with it so on the do you think about this kind of stuff I mean it's exactly what we've been talking about it's like the little details that somehow make you fall in love with the product is that we went from like urban challenge days where love was not part of the conversation, probably.

[670] And to this point where there's human beings and you want them to fall in love with the experience, is that something you're trying to optimize for, trying to think about, like, how do you create an experience that people love?

[671] Absolutely.

[672] I think that's the vision is removing any friction or complexity from getting our users, our writers to where they want to go, making that as simple as possible.

[673] And then, you know, beyond that, just transportation, making, you know, things and, you know, goods get to their destination as seamlessly as possible.

[674] I talked about, you know, a drag and drop experience where I kind of express your intent and then, you know, it just magically happens.

[675] And for our riders, that's what we're trying to get to is you download an app and you, you know, click and car shows up.

[676] It's the same car.

[677] It's very predictable.

[678] It's a safe and high -quality experience, and then it gets you in a very reliable, very convenient, frictionless way to where you want to be.

[679] And along the journey, I think we also want to do little things to delight our users.

[680] Like the ride -sharing companies, because they don't control the experience, I think, they can't make people fall in love necessarily with the experience.

[681] or maybe they haven't put in the effort, but I think if I would just speak to the right sharing experience I currently have, it's just very convenient.

[682] But there's a lot of room for, like, falling in love with it.

[683] Like, we can speak to sort of car companies.

[684] Car companies do this well.

[685] You can fall in love with the car, right?

[686] And be like a loyal car person, like whatever.

[687] Like, I like badass hot rods, I guess 69 Corvette.

[688] And at this point, you know, you can't really, cars are so owning a car is so 20th century, man. But is there something about the Waymo experience where you hope that people will fall in love with it?

[689] Because is that part of it?

[690] Or is it part of, is it just about making a convenient ride, not ride chairing?

[691] I don't know what the right term is, but just a convenient A to B autonomous transport.

[692] Or like, do you want them to fall in love with Waymo?

[693] It may be elaborate a little bit.

[694] I mean, almost like from a business perspective, I'm curious.

[695] Like, how, do you want to be in the background invisible, or do you want to be like a source of joy that's very much in the foreground?

[696] I want to provide the best, most enjoyable transportation solution.

[697] And that means building it, building our product and building our service in a way that people do.

[698] You kind of use in a very seamless, frictionless way in their day -to -day lives.

[699] And I think that does mean, you know, in some way falling in love in that product.

[700] It just kind of becomes part of your routine.

[701] it comes down in my mind to safety, predictability of the experience, and privacy, I think, aspects of it, right?

[702] Our cars, you get the same car, you get very predictable behavior, and that is important, I think, if you're going to use it in your daily life.

[703] Privacy, and when you're in a car, you can do other things.

[704] You're spending a bunch, just another space where you're spending a significant part of your life.

[705] So not having to share it with other people who you don't want to share it with, I think is a very nice property.

[706] Maybe you want to take a phone call or do something else in the vehicle.

[707] And, you know, safety on the quality of the driving as well as the physical safety of not having, you know, to share that ride is, you know, important to a lot of people.

[708] what about the idea that when when there's somebody like a human driving and they do a rolling stop on a stop sign like sometimes like you know you get a Uber or a lift or whatever like human driver and you know they can be a little bit aggressive as as drivers it feels like there is not all aggression is bad now that may be a wrong again 20th century conception of driving maybe it's possible to create a driving experience.

[709] Like, if you're in the back busy doing something, maybe aggression is not a good thing.

[710] It's a very different kind of experience, perhaps.

[711] But it feels like in order to navigate this world, you need to, how do I phrase this?

[712] You need to kind of bend the rules a little bit, or at least I could test the rules.

[713] I don't know what language politicians use to discuss this, but whatever language they use, you flirt with the rules?

[714] I don't know.

[715] But like you sort of have a bit of an aggressive way of driving that asserts your presence in this world, thereby making other vehicles and people respect your presence, and thereby allowing you to sort of navigate through intersections in a timely fashion.

[716] I don't know if any of that made sense, but like how does that fit into the experience of driving autonomously?

[717] It's a lot of this.

[718] You're hitting in a very important point of a number of behavioral components and parameters that make your driving feel assertive and natural, comfortable, Our cars will follow rules, right?

[719] They will do the safest thing possible in all situations.

[720] Let me be clear on that.

[721] But if you think of.

[722] really, really, you know, good drivers.

[723] Just, you know, think about, you know, professional lemon drivers, right?

[724] They will follow the rules.

[725] They're very, very smooth.

[726] And yet, they're very efficient.

[727] And, but they're assertive.

[728] They're comfortable for the people in the vehicle.

[729] They're predictable for the other people outside the vehicle that they share the environment with.

[730] And that's the kind of driver that we want to build.

[731] And you just, you think, if, you know, if maybe there's a sport analogy there, right?

[732] You can do, in many sports, the true professionals are very efficient in their movements, right?

[733] So they don't do like, you know, hectic flailing, right?

[734] They're, you know, smooth and precise, right?

[735] And they get the best results.

[736] So that's the kind of driver that we want to build.

[737] In terms of, you know, aggressiveness, yeah, you can, like, you know, roll through the stop signs.

[738] You can do crazy lane changes.

[739] Typically it doesn't get you to your destination faster.

[740] typically not the safest or most predictable very most comfortable thing to do but there is a way to do both and that that's what we're doing and trying to build the driver that is safe comfortable smooth and predictable yeah that's really interesting distinction I think in the early days of autonomous vehicles the vehicles felt cautious as opposed to efficient and I'm still probably But when I rode in the Waymo, I mean, there was, it was quite assertive.

[741] It moved pretty quickly.

[742] Like, yeah, one of the surprising feelings was that it actually, it went fast.

[743] And it didn't feel like awkwardly cautious than autonomous vehicles.

[744] So I've also programmed autonomous vehicles and everything I've ever built was felt awkwardly, either overly aggressive, okay, especially when it was my code, or, like, awkwardly cautious is the way I would put it.

[745] And Waymo's vehicle felt like assertive and I think efficient is like the right terminology here.

[746] It wasn't, and I also like the professional limo driver.

[747] Because we often think like, you know, an Uber driver or a bus driver or a tax taxi.

[748] This is the funny thing is people think like taxi drivers are professionals.

[749] I mean, it's like, that's like saying I'm a professional walker just because I've been walking all my life.

[750] I think there's an art to it, right?

[751] And if you take it seriously as an art form, then there's a certain way that mastery looks like.

[752] And it's interesting to think about what does mastery look like in driving and perhaps what we associate with like aggressiveness is unnecessary like it's not part of the experience of driving it's like unnecessary fluff that efficiency you could you can be you can create a good driving experience within the rules that's uh i mean you're the first person to tell me this so it's it's kind of interesting.

[753] I need to think about this, but that's exactly what it felt like with Waymo.

[754] I kind of had this intuition.

[755] Maybe it's the Russian thing, I don't know, that you have to break the rules in life to get anywhere.

[756] But maybe, maybe it's possible that that's not the case in driving.

[757] I have to think about that.

[758] But it certainly felt that way on the streets of Phoenix when I was there in Waymo.

[759] That was a very pleasant experience and it wasn't frustrating in that like come on move already kind of feeling it wasn't it that wasn't there yeah i mean that's what that's what we're going after i don't think you have to pick one i think truly good driving it gives you both efficiency assertness but also comfort and predictability and safety uh and you know it's that's what fundamental improvements in the core capabilities truly unlock and you can kind of think of it as, you know, precision and recall trade -off.

[760] You have certain capabilities of your model, and then it's very easy when, you know, you have some curve of precision and recold, you can move things around and can choose your operating point in your trading of precision versus recall, false positives versus false negatives, right?

[761] But then, and, you know, you can tune things on that curve and be kind of more cautious or more aggressive, but then aggressive is bad or, you know, cautious is bad.

[762] But true capabilities come from actually moving the whole curve up, and then New York on a very different plane of those trade -offs.

[763] And that's what we're trying to do here is to move the whole curve up.

[764] Before I forget, let's talk about trucks a little bit.

[765] So I also got a chance to check out some of the Waymo trucks.

[766] I'm not sure if we want to go too much into that space, but it's a fascinating one, so maybe we can mention at least briefly.

[767] You know, Waymo is also now doing autonomous trucking.

[768] and how different, like, philosophically and technically is that whole space of problems?

[769] It's one of our two big products and, you know, commercial applications of our driver, right, right, railing and deliveries.

[770] You know, we have Waymo 1 and Waymo Via, moving people and moving goods.

[771] You know, trucking is an example of moving goods.

[772] We've been working on trucking since 2017.

[773] It is a very interesting space.

[774] And you're a question of how different is it.

[775] It has this really nice property that the first order challenges, like the science, the hard engineering, whether it's hardware or onboard software or offboard software, all of the systems that you build for training your ML models, for evaluating your retirement system.

[776] Like, those fundamentals carry over.

[777] The true challenges of driving, perception, semantic understanding, prediction, decision, making, planning, evaluation, the simulator, ML infrastructure, those carry over.

[778] The data and the application and kind of the domains might be different, but the most difficult problems, all of that carries over between the domains.

[779] So that's very nice.

[780] So that's how we approach it.

[781] We're kind of building investing in the core, the technical core, and then there is specialization of that core technology to different product lines, to different commercial applications.

[782] So just to tease it apart a little bit on trucks, so starting with the hardware, the configuration of the sensors is different, right?

[783] They're different physically, geometrically, different vehicles.

[784] So for example, we have two of the sensors.

[785] of our main laser on the trucks on both sides so that we have not have the blind spots.

[786] Whereas on the JLR IPase, we have one of it sitting at the very top.

[787] But the actual sensors are almost the same, or largely the same.

[788] So all of the investment that over the years we've put into building our custom lighters, custom radars, pulling the whole system together, that carries over very nicely.

[789] Then on the perception side, the fundamental challenges of seeing, understanding the world, whether it's object detection, classification, tracking, semantic understanding, all that carries over.

[790] Yes, there's some specialization.

[791] When you're driving on freeways, range becomes more important.

[792] The domain is a little bit different.

[793] But again, the fundamentals carry over very, very nicely.

[794] Same as you get into prediction or decision making.

[795] The fundamentals of what it takes to predict what other people are going to do, to find the long tail to improve your system in that long tail of behavior prediction and response that carries over right and so on so on so on so I mean that's pretty exciting by the way does way more via include using the the smaller vehicles for transportation of goods that's an interesting distinction so I would say there's three interesting modes of operation so one is moving humans one is moving goods and one is like moving nothing zero occupancy meaning like you're going to the destination you're empty vehicle i mean it's it's the third is the less of me if that's the entire of it it's a less you know exciting from the commercial perspective well well i mean in terms of like if you think about what's inside a vehicle as it's moving because it does you know some significant fraction of the vehicle's movement has to be empty.

[796] I mean, it's kind of fascinating.

[797] Maybe just on that small point, is there different control and, like, policies that are applied for a zero occupancy vehicle, so a vehicle with nothing in it?

[798] Or is it just move as if there is a person inside?

[799] Well, it was with some subtle differences.

[800] As a first order approximation, they're no different.

[801] And if you think about, you know, safety and, you know, confident quality of driving, only part of it, you know, has to do with, you know, the people or the goods inside of the vehicle, right?

[802] But you don't want to be, you know, you want to drive smoothly.

[803] And, you know, as we discussed, not for the purely from the benefit of, you know, whatever you have inside the car, right?

[804] It's also for the benefit of, you know, people outside and kind of fee feeding, fitting naturally and predictably into that whole environment.

[805] right?

[806] So, you know, yes, there's some second order things you can do.

[807] You can change your route and, you know, optimize maybe kind of your fleet things at the fleet scale.

[808] And you would take into account whether, you know, some of your cars are actually, you know, serving a useful trip, whether with people or with goods, whereas, you know, other cars are, you know, driving completely empty to that next valuable trip that they're going to provide.

[809] But those are mostly second order effect.

[810] Okay, cool.

[811] So Phoenix is, uh, is an incredible place.

[812] And what you've announced in Phoenix is, uh, it's kind of amazing.

[813] But, you know, that's just like one city.

[814] How do you take over the world?

[815] Uh, I mean, I'm asking for a friend.

[816] One, once, one step at a time.

[817] Pinky, is that a cartoon pinky in the brain?

[818] Yeah.

[819] Okay.

[820] But, you know, but, you know, gradually is a true answer.

[821] So I think the heart of your question is, can you ask a better question than I ask?

[822] You're asking a great question.

[823] Answer that one.

[824] I'm just going to, you know, phrase it in the terms that I want to answer.

[825] Perfect.

[826] This is exactly right.

[827] Brilliant.

[828] Please.

[829] You know, where are we today?

[830] And, you know, what happens next?

[831] And what does it take to go beyond Phoenix and what does it take to get this technology to more places and more people around the world, right?

[832] So our next big area of focus is exactly that larger scale commercialization and scaling up.

[833] If I think about the main And, you know, Phoenix gives us that platform and gives us that foundation of upon which we can build.

[834] And it's, there are few really challenging aspects of this whole problem that you have to pull together in order to build a technology, in order to deploy it into the field, to go from a driver car to a fleet of cars that are providing a service and then all the way to, you know, commercialization.

[835] So, and, you know, this is what we have in Phoenix.

[836] We've taken the technology from a proof point to an actual deployment and have taken our driver, you know, from, you know, one car to a fleet that can provide a service.

[837] Beyond that, if I think about what it will take to scale up and deploy in, you know, more places with more customers, I tend to think about three main dimensions, three main axes of scale.

[838] One is the core technology, you know, the hardware and software, core capabilities of our driver.

[839] The second dimension is evaluation and deployment.

[840] And the third one is, you know, the third one is, you know, It's just the product, commercial, and operational excellence.

[841] So you can talk a bit about where we are along each one of those three dimensions about where we are today and what will happen next.

[842] On the core technology, the hardware and software, together comprise of driver, we obviously have that foundation that is providing fully driverless trips.

[843] to our customers as we speak, in fact.

[844] And we've learned a tremendous amount from that.

[845] So now what we're doing is we are incorporating all those lessons into some pretty fundamental improvements in our core technology, both on the hardware side and on the software side, to build a more general, more robust solution that then will enable us to massively scale beyond Phoenix.

[846] So on the hardware side, all of those, lessons are now incorporated into this fifth generation hardware platform that is being deployed right now.

[847] And that's the platform.

[848] The fourth generation, the thing that we have right now driving in Phoenix, it's good enough to operate fully driverlessly, you know, night and day, you know, various speeds and various conditions.

[849] But the fifth generation is the platform upon which we want to go to massive scale.

[850] We, in turn, we've really made qualitative improvements in terms of the capability of the system, the simplicity of the architecture, the reliability of the redundancy.

[851] It is designed to be manufacturable at very large scale and provides the right unit economics.

[852] So that's the next big step for us on the hardware side.

[853] That's already there for scale, the version 5.

[854] That's right.

[855] Is that coincidence or should we look into a conspiracy theory that it's the same version as the pixel phone?

[856] Is that what's the hardware?

[857] I can neither.

[858] confirm or deny Lux.

[859] All right, cool.

[860] So, sorry, so that's the, okay, that's that axis.

[861] What else?

[862] So similarly, you know, hardware is a very discrete jump, but similar to the, uh, that, to how we're making that change from the fourth generation hardware to the fifth, we're making similar improvements on the software side to make it more, you know, robust and more general and allow us to, you know, quickly, uh, scale beyond Phoenix.

[863] So that, that's the first dimension of core technology.

[864] The second dimension is evaluation and deployment.

[865] Now, how do you, uh, measure your system.

[866] How do you evaluate it?

[867] How do you build a release and deployment process where, you know, with confidence, you can, you know, regularly release new versions of your driver into a fleet?

[868] How do you get good at it so that it is not, you know, a huge tax on your researchers and engineers that, you know, so you can, how do you build all these, you know, processes, the frameworks, the simulation, the evaluation, the data science, the validation, so that, you know, people can focus on improving the system and kind of the releases just go out the door and get deployed across the fleet.

[869] So we've gotten really good at that in Phoenix.

[870] That's been a tremendously difficult problem.

[871] But that's what we have in Phoenix right now.

[872] That gives us that foundation.

[873] And now we're working on incorporating all the lessons that we've learned to make it more efficient to go to new places and scale up and just kind of, you know, stamp things out.

[874] So that's that second dimension of evaluation and deployment.

[875] And the third dimension is product, commercial, and operational excellence.

[876] And again, Phoenix there is providing an incredibly valuable platform.

[877] That's why we're doing things end -to -end in Phoenix.

[878] We're learning, as we discussed a little earlier today, tremendous amount of really valuable lessons from our users getting really incredible feedback.

[879] And we'll continue to iterate on that and incorporate all those lessons and to making our product even better and more convenient for our users.

[880] So you're converting this whole process of Phoenix in Phoenix into something that could be copy and pasted elsewhere.

[881] So like perhaps you didn't think of it that way when you were doing the experimentation Phoenix.

[882] So how long did, basically, you can correct me, but you've, I mean, it's still early days, but you've taken a full journey in Phoenix, right?

[883] as you were saying, of, like, what it takes to basically automate.

[884] I mean, it's not the entirety of Phoenix, right?

[885] But I imagine it can encompass the entirety of Phoenix.

[886] That's some near -term date.

[887] But that's not even perhaps important, as long as it's a large enough geographic area.

[888] So how copy -pasteable is that process currently?

[889] and how like you know like when you copy and paste in Google docs I think or in Word you can like apply source formatting or apply destination formatting so when you copy and paste the Phoenix into like say Boston how do you apply the destination formatting like how much of the core of the entire process of bringing an actual public transportation, autonomous transportation service to a city is there in Phoenix that you understand enough to copy and paste into Boston, or wherever.

[890] So we're not quite there yet.

[891] We're not at a point where we're kind of massively copy and pasting all over the place.

[892] But Phoenix, what we did in Phoenix, and we very intentionally have chosen Phoenix as our, our first full deployment area, you know, exactly for that a reason to kind of tease the problem apart, look at each dimension, focus on the fundamentals of complexity and derisking those dimensions, and then bringing the entire thing together to get all the way, and force ourselves to learn all those hard lessons on this chronology, hardware and software, on the evaluation deployment, on operating a service, operating a business, using, actually serving our customers all the way so that we're fully informed about the most difficult, most important challenges to get us to that next step of massive copy and pasting, as you said.

[893] And that's what we're doing right now.

[894] We're incorporating all those things that we learned into that next system that then will allow us to kind of copy and paste all over the place and to massively scale to more users and more locations.

[895] I mean, you know, just talked a little bit about, you know, what does that mean along those different dimensions?

[896] So on the hardware side, for example, again, it's that switch from the fourth to the fifth generation.

[897] And the fifth generation is designed to kind of have that property.

[898] Can you say what other cities you're thinking about?

[899] Like, I'm thinking about, sorry, we're in San Francisco now.

[900] I thought I want to move to San Francisco, but I'm thinking about moving to Austin.

[901] I don't know why.

[902] People are not being very nice about San Francisco currently.

[903] Maybe it's a small, maybe it's in vogue right now.

[904] But Austin seems, I visited there and it was, I was in a Walmart.

[905] It's funny, these moments, like, turn your life.

[906] There's this very nice woman with kind eyes, just, like, stopped and said, he looked so handsome in that tie, honey, to me. This has never happened to me in my life, but just the sweetness of this woman.

[907] And it's something I've never experienced, certainly in the streets of Boston, but even in San Francisco where people wouldn't, that's just not how they speak or think.

[908] I don't know.

[909] There's a warmth to Austin that love.

[910] And since Waymo does have a little bit of a history there, is that a possibility?

[911] Is this your version of asking the question of like, you know, Dimitri, I know you can't share your commercial and deployment roadmap, but I'm thinking about moving to San Francisco of Austin, like, you know, blink twice if you think I should move to him.

[912] Yeah, that's true.

[913] That's true.

[914] You got me. We, you know, we've been testing in all over the place.

[915] I think we've been testing more in, you know, 25 cities.

[916] We drive in San Francisco.

[917] We drive in, you know, Michigan for snow.

[918] We are doing significant amount of testing in the Bay Area and including San Francisco.

[919] But it's not like, because we're talking about the very different thing, which is like a full -on, large geographic area public service.

[920] You can't share.

[921] Okay.

[922] what about Moscow is that when is that happening take on yandex i'm not paying attention of those folks they're doing you know there's there's a lot of fun i mean maybe that as a way of a question you didn't speak to sort of like policy or like is there tricky things with government and so on like is there other friction that you've encountered except sort of technological friction of solving this very difficult problem?

[923] Is there other stuff that you have to overcome when deploying a public service in a city?

[924] That's interesting.

[925] It's very important.

[926] So we put significant effort in creating those partnerships and those relationships with governments at all levels, local governments, municipalities, you know, state level, federal level, we've been engaged in very deep conversations from the earliest days of our projects whenever at all of these levels, you know, whenever we go to test or, you know, operate in a new area, you know, we always lead with the conversation with the local officials.

[927] But the result of that, that investment is that, no, it's not challenges we have to overcome, but it is a very important that we continue to have this conversation oh yeah i love politicians do okay uh so mr elan musk said that uh lidar is a crutch what are your thoughts i wouldn't characterize it exactly that way uh i know i think light lighters is very important uh it is a key sensor uh that you know we use just like other modalities right as we discussed our cars use cameras uh lighters and radars they are all very important they are at kind of the physical level they are very different they have very different you know physical characteristics cameras are passive widars and radars are active use different wavelengths so that means they complement each other very nicely and and they together combined they can be used to build a much safer and much more capable system.

[928] So to me, it's more of a question, you know, why the heck would you handicap yourself and not use one or more of those sensing modalities when they, you know, undoubtedly just make your system more capable and safer.

[929] Now, it, you know, what might.

[930] make sense for one product or one business might not make sense for another one.

[931] So if you're talking about driver assist technologies, you make certain design decisions and you make certain tradeoffs and you make different ones if you are building a driver that deploy in fully driverless vehicles.

[932] And lighter specifically when this question comes up, I, you know, typically the criticisms that I hear, or, you know, the counterpoints that cost and aesthetics.

[933] And I don't find either of those, honestly, very compelling.

[934] So on the cost side, there's nothing fundamentally prohibitive about, you know, the cost of riders.

[935] You know, radars used to be very expensive before people started, you know, before people made certain advances in technology and, you know, started to manufacture them at massive scale and deploying them in vehicles, right, similar with lighters.

[936] And this is where the lighters that we have on our cars, especially the fifth generation, we've been able to make some pretty qualitative, discontinuous jumps in terms of the fundamental technology that allow us to manufacture those things at very significant scale and add a fraction of the cost of both our previous generation, as well as a fraction of the cost of what might be available on the market, you know, off the shelf right now.

[937] And, you know, that improvement will continue.

[938] So I, I think, you know, cost is not a real issue.

[939] Second one is, you know, aesthetics.

[940] You know, I don't think that's, you know, a real issue either.

[941] Udius and I, the beholder.

[942] You can make LIDAR sexy again.

[943] I think you're exactly right.

[944] I think it is sexy.

[945] Like, honestly, I think form and function.

[946] Yeah, I always thought, you know, I was actually, somebody brought this up to me. I mean, all forms of LIDAR, even, like, the ones that are, like, big, you can make look, I mean, it can make look beautiful.

[947] Like, there's no sense in which you can't integrate into design.

[948] Like, there's all kinds of awesome designs.

[949] I don't think small and humble is beautiful.

[950] It could be, like, you know, brutalism or, like, it could be, like, harsh corners.

[951] I mean, like I said, like hot rods.

[952] Like, I don't like, I don't necessarily like, like, oh, man, I'm going to start so much controversy with this.

[953] I don't like Porsches, okay, the Porsche 9 -11, like everyone says, oh, the most beautiful, no, it's like, it's like a baby car.

[954] It doesn't make any sense.

[955] But everyone, it's beauties and eye at the beholder, you're already looking at me like, what's this kid talking about?

[956] I'm happy to talk about, you're digging your own hole.

[957] The form and function and my take on the beauty.

[958] beauty of the hardware that we put on our vehicles, you know, I will not comment on your Porsche monologue.

[959] Okay.

[960] All right.

[961] So, but aesthetics, fine.

[962] But there's an underlying, like, philosophical question behind the kind of lighter question is, like, how much of the problem can be solved with computer vision, with machine learning.

[963] So I think without sort of disagreements and so on.

[964] It's nice to put it on the spectrum because Waymo is doing a lot of machine learning as well.

[965] It's interesting to think how much of driving, if we look at five years, 10 years, 50 years down the road, can be learned in an almost more and more and more end -to -end way.

[966] If you look at what Tesla is doing as a machine learning problem, they're doing a multitask learning thing where it's just they break up driving into a bunch of learning tasks.

[967] they have a single neural network and they're just collecting huge amounts of data that's training that.

[968] I've recently hung out with George Hatz.

[969] I don't know if you know George.

[970] I love him so much.

[971] He's just an entertaining human being.

[972] We were off mic talking about Hunter S. Thompson.

[973] He's the Hunter S. Thompson of autonomous driving.

[974] Okay.

[975] So he, I didn't realize this with comma AI, but they're like really trying to do end to end.

[976] They're, like, looking at the machine learning problem, they're, really not doing multitask learning, but it's computing the drivable area as a machine learning task and hoping that down the line this level two system that's driver assistance will eventually lead to allowing you to have a fully autonomous vehicle.

[977] Okay, there's an underlying deep philosophical question there, technical question of how much of driving can be learned.

[978] So LIDAR is an effective tool today for actually deploying a successful service in Phoenix, right, that's safe, that's reliable, et cetera, et cetera.

[979] But the question, and I'm not saying you can't do machine learning on LIDAR, but the question is that like how much of driving can be learned eventually?

[980] Can we do fully autonomous that's learned?

[981] Yeah.

[982] You know, learning is all over the place and plays a cue role in every part of our system.

[983] I, as you say, said, I would decouple the sensing modalities from the, you know, ML and the software parts of it.

[984] Lider, radar, cameras, it's all machine learning.

[985] All of the object detection classification, of course, like, well, that's what, you know, these modern deep nets and continents are you very good at.

[986] You feed them raw data, massive amounts of raw data.

[987] And, you know, that's actually what our custom -built lighters and raters are really good at.

[988] And radars, they don't just give you point estimates of, you know, objects in space, they give you raw, like, physical observations.

[989] And then you take all of that raw information, you know, whether it's colors of the pixels, whether it's, you know, lighters returns and some auxiliary information.

[990] It's not just distance, right?

[991] And, you know, angle and distance is much richer information that you get from those returns, plus really rich information from the radars.

[992] You fuse it all together and you feed it into those massive ML models that then, you know, lead to the best results in terms of, you know, object detection, classification, state estimation.

[993] So there's a, sorry to interrupt, but there is a fusion.

[994] I mean, that's something that people didn't do for a very long time, which is like at the sensor fusion level, I guess, like early on fusing the information together, whether so that the sensory information that the vehicle receives from the different modalities or even from different cameras is combined before it is fed into the machine learning models.

[995] Yes, I think this is one of the trends.

[996] You're seeing more of that.

[997] You mentioned N -10.

[998] There's different interpretation of N -10.

[999] There is kind of the purest interpretation of I'm going to like have one model that goes from raw sensor data to like, you know, steering torque and, you know, gas brakes.

[1000] That, you know, that's too much.

[1001] I don't think that's the right way to do it.

[1002] There's more, you know, smaller versions of end to end where you're, you know, kind of doing more end -to -end learning or core training or deep propagation of kind of signals back and forth across the different stages of your system.

[1003] There's, you know, really good ways it gets into some fairly complex design.

[1004] choices, where in one hand, you want modularity and decompositability, the composability of your system.

[1005] But on the other hand, you don't want to create interfaces that are too narrow or too brittle, too engineered, where you're giving up on the generality of a solution, or you're unable to properly propagate signal, you know, reach signal forward and losses and, you know, back so you can optimize the whole system jointly.

[1006] So I would decouple, and I guess what you're seeing in terms of the fusion of the sensing data from different modalities, as well as kind of fusion in the temporal level going more from, you know, frame by frame, where, you know, you would have one net that would do frame by frame detection and camera.

[1007] And then, you know, something that does frame by frame and lighter and then radar.

[1008] And then you fuse it, you know, in a weaker engineered way later.

[1009] The field over the last, you know, decade has been evolving in more kind of joint fusion, more end -to -end models that are, you know, solving some of these tasks, you know, jointly.

[1010] And there's tremendous power in that.

[1011] And, you know, that's the progression that, you kind of our stack has been on as well.

[1012] Now, so I would decouple the sensing and how that information is used from the role of ML and the entire stack.

[1013] And, you know, I guess it's, I, there's straight -ups and, you know, modularity and how do you inject inductive bias into your system, right?

[1014] This is, there's tremendous power in being able to do that.

[1015] So, you know, we have, there's no part of our system that is not heavily, that does not heavily, you know, leverage data -driven development or, you know, state of the art -mell.

[1016] But there's mapping, there's a simulator, or there's perception, you know, object -level, perception, whether it's semantic understanding, prediction, decision -making, you know, so forth and so on.

[1017] It's, and of course, object detection and classification, like, you're finding pedestrians and cars and cyclists and, you know, cones and signs and vegetation and being very good at estimating kind of detection, classification, and state estimation, there's just stable stakes.

[1018] Like that's step zero of this whole stack.

[1019] You can be incredibly good at that, whether you use cameras or light as a radar, but that's just, you know, that's stable stakes.

[1020] That's just step zero.

[1021] Beyond that, you get into the really interesting challenges of semantic understanding, the perception level.

[1022] You get into scene level reasoning.

[1023] You get into very deep problems that have to do with prediction and joint prediction and interaction, so on an interaction between all of the actors in the environment, pedestrian, cyclists, other cars, and you get into decision -making, right?

[1024] So how do you build a lot of systems?

[1025] So we Leverage ML very heavily in all of these components.

[1026] I do believe that the best results you achieve by kind of using a hybrid approach and having different types of ML, having different models with different degrees of inductive bias that you can have, and combining kind of model free approaches with some model -based approaches and some rule -based, physics -based systems.

[1027] So one example I can give you.

[1028] you is traffic lights.

[1029] There's a problem of the detection of traffic light state, and obviously that's a great problem for computer vision.

[1030] Confidence, that's their bread and butter.

[1031] That's how you build that.

[1032] But then the interpretation of a traffic light, that you're going to need to learn that.

[1033] You don't need to build some complex ML model that infers with some precision and recall that red means stop.

[1034] It's a very clear engineered signal with very clear semantics.

[1035] So you want to induce that bias.

[1036] Like how you induce that bias and that, whether it's a constraint or a cost function in your stack, but it is important to be able to inject that clear semantic signal into your stack.

[1037] And that's what we do.

[1038] And but then the question of like, and that's when you apply it to yourself, when you are making decisions whether you want to stop for a red light, you know, or not.

[1039] But if you think about how other people treat traffic lights, we're back to the ML version of that.

[1040] You know they're supposed to stop for a red light, but that does mean they will.

[1041] So then you're back in the like very heavy ML domain where you're picking up on like very subtle keys about, you know, they have to do with the behavior of objects, pedestrians, cyclists, cars, and the whole, you know, the entire configuration of the scene that allow you to make accurate predictions on whether they will, in fact, stop or run a red light.

[1042] So it sounds like already for Waymo, like machine learning is a huge part of the stack.

[1043] So it's a huge part of like not just, so obviously the first level zero or whatever you said, which is like just the object detection of things that, you know, with know that machine learning can do, but also starting to do prediction, behavior and so on to model the what other or the other parties.

[1044] in the scene, entities in the team are going to do.

[1045] So machine learning is more and more playing a role in that as well.

[1046] Of course.

[1047] Oh, absolutely.

[1048] I think we've been going back to the earliest days like DARPA Grand Challenge.

[1049] And team was leveraging machine learning.

[1050] I was like pre -image nut and it was very different type of ML.

[1051] And I think actually it was before my time, but the Stanford team during the Grand Challenge had a very interesting machine learned system that would use lighter.

[1052] and camera.

[1053] We've been driving in the desert, and we had built the model where it would kind of extend the range of free space reasoning.

[1054] We'd get a clear signal from lighter.

[1055] And then it had a model that said, hey, like this stuff in camera kind of sort of looks like this stuff in lighter.

[1056] And I know this stuff that I've seen in lighter.

[1057] I'm very confident that's free space.

[1058] So let me extend that free space zone into the camera range that would allow the vehicle to drive faster.

[1059] And then we've been building on top of that and kind of staying and pushing the state of the art in ML, in all kinds of different ML over the years.

[1060] And in fact, from the earlier days, I think 2010 is probably the year where Google, maybe 2011 probably, got pretty heavily involved in machine learning, kind of deep nuts.

[1061] And at that time, it's probably the only company that was very heavily investing in kind of state -of -the -art ML and self -driving cars.

[1062] And they go hand -in -hand.

[1063] and we've been on that journey ever since.

[1064] We're doing, pushing a lot of these areas in terms of research at Waymo, and we collaborate very heavily with the researchers in alphabet.

[1065] And like all kinds of ML, you know, supervised ML, unsupervised ML.

[1066] You know, published some interesting research papers in the space, especially recently.

[1067] It's just a super super active.

[1068] Yeah, super super active.

[1069] Of course, there's, you know, kind of like more maturist.

[1070] stuff like, you know, confinets for, you know, object detection.

[1071] But there's some really interesting, really active work that's happening in, kind of more, you know, and bigger models in, you know, models that have more structure to them, you know, not just, you know, large bitmaps and reason about temporal sequences.

[1072] And some of the interesting breakthroughs that you, you know, we've seen in language models, right?

[1073] You know, transformers, you know, GPT, you know, 3, and friends, there's some really interesting applications of some of the core breakthroughs to those problems of, you know, behavior prediction as well as, you know, decision making and planning, right?

[1074] You can think about it, kind of the behavior, how, you know, the path, the trajectories, the how people drive, and they have kind of a share a lot of the fundamental structure, you know, this problem.

[1075] There's, you know, sequential, you know, nature.

[1076] There's a lot of structure in this representation.

[1077] There is a strong locality, kind of like in sentences, you know, words that follow each other.

[1078] They're strongly connected, but there are also kind of larger context that doesn't have that locality and you also see that on driving, right?

[1079] What, you know, is happening in the scene as a whole has very strong implications on, you know, the kind of the next step in that sequence where whether you're predicting what other people are going to do, whether you're making your own decisions, or whether in a simulator, you're building generative models of, you know, humans, walking, cyclist riding, and other cars driving.

[1080] Oh, that's all really fascinating.

[1081] It's fascinating to think that transform our models and all the breakthroughs in language and nLP that might be applicable to like driving at the higher level at the behavior level that's kind of fascinating let me ask about pesculative creatures called pedestrians and cyclists they seem so humans are a problem if we can get rid of them i would but unfortunately they're also a source of joy and love and beauty so let's keep them around they're also our customers for your perspective yes yes for sure some money very good um but uh i don't even know where i was going oh yes pedestrians and cyclists i you know they're a fascinating injection into the system of uncertainty of um of like a game theoretic dance of what to do and and also uh they have perceptions of their own and they can tweet about your product, so you don't want to run them over from that perspective.

[1082] I mean, I don't know.

[1083] I'm joking a lot, but I think in seriousness, like, you know, pedestrians are complicated computer vision problem, a complicated behavioral problem.

[1084] Is there something interesting you could say about what you've learned from a machine learning perspective, from also an autonomous vehicle, and a product perspective about just interacting with the humans in this world?

[1085] Yeah.

[1086] Just, you know, to state on the record, we care deeply about the safety of pedestrians, you know, even the ones that don't have Twitter accounts.

[1087] Thank you.

[1088] All right.

[1089] All right, cool.

[1090] Not me. But yes, I'm glad I'm glad somebody does.

[1091] Okay.

[1092] But, you know, in all seriousness, safety of vulnerable road users, pedestrians or cyclists, is one of our highest priorities.

[1093] We do a tremendous amount of testing and validation and put a very significant emphasis on the capabilities of our systems that have to do with safety around those unprotected vulnerable road users.

[1094] Cars just discussed earlier in Phoenix, we have completely empty cars, completely driverless cars, driving in this very large area.

[1095] And some people use them to go to schools.

[1096] I will drive through school zones.

[1097] So kids are kind of the very special class of those vulnerable user, road users, right?

[1098] You want to be super, super, super safe and super, super, super cautious around those.

[1099] So we take it very, very, very seriously.

[1100] And, you know, what does it take to be good at it?

[1101] You know, an incredible amount of performance across your whole stack.

[1102] You know, starts with hardware.

[1103] and again, you want to use all sensing modalities available to you.

[1104] Imagine driving on a residential road at night and kind of making a turn and you don't have headlights covering some part of the space and like a kid might run out.

[1105] And lighters are amazing at that.

[1106] They see just as well in complete darkness as they do during the day, right?

[1107] So again, it gives you that extra margin in terms of capability and performance and safety and quality.

[1108] And in fact, we oftentimes, in these kinds of situations, we have our system detect something, in some cases even earlier than our train operators in the car might do, especially in conditions like very dark nights.

[1109] So starts with sensing.

[1110] Then perception has to be incredibly good.

[1111] And you have to be very, very good at detecting pedestrians in all kinds of situations and all kinds of environments, including people in weird poses, people kind of running around and being partially occluded.

[1112] So, you know, that's stuff number one.

[1113] Then you have to have in very high accuracy and very low latency in terms of your reactions to, you know, what, you know, these actors might do.

[1114] And we've put a tremendous amount of engineering and tremendous amount of validation into make sure our system performs properly.

[1115] And oftentimes it does require a very strong reaction to do the safe thing.

[1116] And we actually see a lot of cases like that.

[1117] That's the long tail of really rare, you know, really crazy events that contribute to the safety around pedestrians.

[1118] One example that comes to mind that we actually happened in Phoenix where we were driving along.

[1119] And I think it was a 45 -mine.

[1120] all per hour road, so you know, pretty high -speed traffic.

[1121] And there was a sidewalk next to it, and there was a cyclist on the sidewalk.

[1122] And as we were in the right lane, right next to the sidewalk, so as we got close to the cyclist on the sidewalk, it was a woman, she tripped and fell.

[1123] Just, you know, fell right into the path of our vehicle.

[1124] And our car, you know, this was actually with a test driver, our test drivers, did exactly the right thing.

[1125] they kind of reacted and came to stop.

[1126] It requires both very strong steering and strong application of the brake.

[1127] And then we simulated what our system would have done in that situation.

[1128] And it did exactly the same thing.

[1129] And that speaks to all of those components of really good state estimation and tracking.

[1130] And imagine a person on a bike and they're falling over and they're doing that right in front of you.

[1131] So you have to be really like things are changing.

[1132] The appearance of that whole thing is changing.

[1133] And a person goes one way.

[1134] They're falling on the road.

[1135] they're, you know, being flat on the ground in front of you, you know, the bike goes flying the other direction, like the two objects that used to be one, they're now, you know, are splitting apart, and the car has to, like, detect all of that, like, milliseconds matter, and it doesn't, you know, it's not good enough to just break.

[1136] You have to, like, steer and break, and there's traffic around you.

[1137] So, like, it all has to come together, and it was really great to see in this case, in other cases like that, that we're actually seeing in the wild, that our system is, you know, performing exactly the way that we would have liked.

[1138] And as a. able to, you know, avoid collisions like this.

[1139] It's such an exciting space for robotics, like in that split second to make decisions of life and death.

[1140] I don't know.

[1141] The stakes are high in the sense, but it's also beautiful that, for somebody who loves artificial intelligence, the possibility that an AI system might be able to save a human life, that's kind of exciting as a problem, like to wake up.

[1142] It's terrifying probably for an engineer to wake up.

[1143] and to think about, but it's also exciting, because it's like it's in your hands.

[1144] Let me try to ask a question that's often brought up about autonomous vehicles, and it might be fun to see if you have anything interesting to say, which is about the trolley problem.

[1145] So a trolley problem is an interesting philosophical construct that highlights, and there's many others like it, of the difficult ethical decisions that we humans, have before us in this complicated world.

[1146] So specifically is the choice between if you were forced to choose to kill a group X of people versus a group Y of people, like one person.

[1147] If you did nothing, you would kill one person, but if you would kill five people and if you decide to swerve out of the way, you would only kill one person.

[1148] Do you do nothing or you choose to do something?

[1149] and you can construct all kinds of sort of ethical experiments of this kind that I think at least on a positive note inspire you to think about like introspect what are the the physics of our morality and there's usually not good answers there I think people love it because it's just an exciting thing to think about I think people who build autonomous vehicles usually roll their eyes and because this is not, this one as constructed, this like literally never comes up in reality.

[1150] You never have to choose between killing one or like one of two groups of people.

[1151] But I wonder if you can speak to, is there something interesting to use an engineer of autonomous vehicles that's within the trolley problem or maybe more generally, are there, difficult ethical decisions that you find that algorithm must make.

[1152] On the specific version of the trial problem, which one would you do if you're driving?

[1153] The question itself is a profound question because we humans ourselves cannot answer it, and that's the very point.

[1154] I will kill both.

[1155] Yeah, humans, I think you're exactly right, and that humans are not particularly good.

[1156] I think they kind of phrased as a, like, what would a computer do?

[1157] but humans are not very good.

[1158] And actually, oftentimes, I think that freezing and kind of not doing anything because you've taken a few extra milliseconds to just process and then you end up doing the worst of the possible outcomes, right?

[1159] So I do think that, as you've pointed out, it can be a bit of a distraction and it can be a bit of a kind of a red herring.

[1160] I think it's an interesting discussion in the realm of philosophy, right?

[1161] But in terms of how that affects the actual engineering and deployment of self -driving vehicles.

[1162] It's not how you go about building a system.

[1163] We've talked about how you engineer a system, how you go about evaluating the different components and the safety of the entire thing.

[1164] How do you kind of inject the various model -based safety -based arguments?

[1165] And yes, you reason it parts of the system.

[1166] You know, you reason about the probability of a collision, the severity of that collision.

[1167] right?

[1168] And that is incorporated and there's, you know, you have to properly reason about the uncertainty that flows through the system, right?

[1169] So, you know, those, you know, factors definitely play a role in how the cars then behave, but they isn't to be more of like the emergent behavior.

[1170] And what you see, like, you're absolutely right, that these, you know, clear theoretical problems that they, you know, you don't accord that in system.

[1171] And really kind of being back to our previous discussion, like, what, you know, what, what, you know, which one do you choose?

[1172] Well, you know, oftentimes, like, you made a mistake earlier.

[1173] Like, it shouldn't be in that situation in the first place, right?

[1174] And in reality, the system comes up.

[1175] If you build a very good, safe and capable driver, you have enough, you know, clues in the environment that you drive defensively so you don't put yourself in that situation, right?

[1176] And again, you know, it has, you know, if you go back to that analogy of, you know, precision and recall, like, okay, you can make a, you know, very hard trade -off off, but like neither answer is really good.

[1177] But what instead you focus on is kind of moving the whole curve up and then you focus on building the right capability and the right defensive driving so that, you know, you don't put yourself in the situation like this.

[1178] I don't know if you have a good answer for this, but people love it when I ask this question about books.

[1179] Are there books in your life that you've enjoyed philosophical, fiction, technical, that had a big impact in you as an engineer or as a human being, you know, everything from science fiction to a favorite textbook.

[1180] Is there three books that stand out that you can think of?

[1181] Three books.

[1182] So I would, you know, that impacted me. I would say, and this one is, you probably know it well, but not generally well known.

[1183] I think in the U .S. are kind of internationally, The Master and Margarita.

[1184] It's one of actually my favorite books.

[1185] It is, you know, by our, Russian, it's a novel by Russian author, Mikhail Bulgakov.

[1186] And it's just, it's a great book.

[1187] You know, it's one of those books that you can, like, reread your entire life.

[1188] And it's very accessible.

[1189] You can read it as a kid.

[1190] And, like, it's, you know, the plot is interesting.

[1191] It's, you know, the devil, you know, visiting the Soviet Union.

[1192] And, you know, like, you read it, reread it at different stages of your life.

[1193] And you, you know, you enjoy it for different, very different reasons.

[1194] And you could find, like, deeper and deeper meaning.

[1195] And, you know, kind of affected, you know, had a, definitely had an, like, imprint on me, but, you know, mostly from the, probably kind of the cultural stylistic aspect.

[1196] Like, it makes you the amount of those books that, you know, is good and makes you think, but also has like this really, you know, silly, quirky, dark sense of, you know, humor.

[1197] Yeah, it captures the Russian soul more than maybe, perhaps many other books.

[1198] On that, like, slight note, just out of curiosity, one of the saddest things is I've read that book in English, did you, by chance, read it in English?

[1199] or in Russian?

[1200] In Russian, only in Russian.

[1201] And actually, that is a question I had.

[1202] I kind of posed to myself every once in a while.

[1203] I wonder how well it translates, if it translates at all.

[1204] And there's the language aspect of it, and then there's the cultural aspect.

[1205] So actually, I'm not sure if either of those would work well in English.

[1206] Now, I forget their names, but so when the COVID lifts a little bit, I'm traveling to Paris for several reasons.

[1207] One is just I've never been to Paris.

[1208] I want to go to Paris.

[1209] But there's the most famous translators of Dostoe, of most of Russian literature, live there.

[1210] There's a couple.

[1211] They're famous, a man and a woman, and I'm going to sort of have a series of conversations with them.

[1212] And in preparation for that, I'm starting to read Dostoevsky in Russian.

[1213] So I'm really embarrassed to say that I read everything I've read of Russian literature of, like, serious depth has been in English.

[1214] even though I can also read I mean obviously in Russian but for some reason it seemed in the optimization of life it seemed the improper decision to do it to read in Russian like I don't need to I need to think in English not in Russian but now I'm changing my mind on that and so the question of how well it translate is a really fundamental one like even with Dostoevsky so for what I understand does the steveskin translates easier uh others don't as much obviously the poetry doesn't translate as well i'm also the the music of big fan of latimer wosotsky he doesn't obviously translate well people have tried but master i don't know i don't know about that one i just know it in english you know it's fun fun as hell in english so uh so but it's a curious question and i want to study rigorously from both the machine learning aspect and also because I want to do a couple of interviews in Russia that I'm still unsure of how to properly conduct an interview across a language barrier.

[1215] It's a fascinating question that ultimately communicates to an American audience.

[1216] There's a few Russian people that I think are truly special human beings.

[1217] And I feel I sometimes encounter this with some incredible scientists and maybe you encounter this as well at some point in your life that it feels like because of the language barrier, their ideas are lost to history.

[1218] It's a sad thing.

[1219] I think about like Chinese scientists or even authors that like that we don't, in an English -speaking world, don't get to appreciate some like the depth of the culture because it's lost in translation.

[1220] And I feel like I would love to show that to the world.

[1221] Like, I'm just some idiot, but because I have this, like, at least some semblance of skill in speaking Russian, I feel like, and I know how to record stuff on a video camera, I feel like I want to catch, like, Gregory Perlman, who's a mathematician, I'm not sure if you're familiar with him.

[1222] I want to talk to him, like, he's a fascinating mind, and to bring him to a wider audience in English speaking, it'll be fascinating.

[1223] but that requires to be rigorous about this question of how well Bulgakov translates.

[1224] I mean, I know it's a silly concept, but it's a fundamental one, because how do you translate?

[1225] And that's the thing that Google Translate is also facing as a more machine learning problem.

[1226] But I wonder as a more bigger problem for AI, how do we capture the magic that's there in the language?

[1227] thing that's a really interesting, really challenging problem.

[1228] If you do read it, master in margaria in English, sorry, in Russian and be curious, just get your opinion.

[1229] And I think part of it is language, but part of it's just, you know, centuries of culture that, you know, the cultures are different.

[1230] So it's hard to connect that.

[1231] Okay, so that was my first one, right?

[1232] You had two more.

[1233] The second one I would probably pick the science fiction by the Strogaski brothers.

[1234] You know, it's up there with, you know, Isaac Asimov and, you know, Ray Bradbury and, you know, company.

[1235] The Strogotsky brothers kind of appealed more to me. I think more, I made more of an impression on me growing up.

[1236] I apologize if I'm showing my complete ignorance.

[1237] I'm so weak on sci -fi.

[1238] What do they write?

[1239] Oh, roadside picnic.

[1240] Hard to be a god.

[1241] Beetle in an End Hill Monday starts on Saturday it's not just science fiction it's also like has very interesting interpersonal and societal questions and some of the language is just completely hilarious that's the one that's the one that's interesting Monday starts on Saturday so I need to read okay oh boy you put that in a category of science fiction That one is, I mean, this was more of a silly, you know, humorous work.

[1242] I mean, there is kind of, you know, it's profound too, right?

[1243] It's science fiction, right?

[1244] It's about, you know, this research institute and like it's, it has deep parallels to like serious research, but the setting, of course, is that they're working on, you know, magic, right?

[1245] And there's a lot of, so I, I, and that's their style, right?

[1246] And, you know, other books are very different, right?

[1247] You know, hard to be a god, right?

[1248] It's about kind of this higher society being injected into this primitive world and how they operate there on some of the very deep ethical, you know, questions there, right?

[1249] And like they've got this full spectrum.

[1250] Some, you know, more about kind of more adventure style.

[1251] But like I enjoy all of their books.

[1252] There's just, you know, probably a couple.

[1253] Actually, one, I think that they consider their most important work.

[1254] I think it's the snail on an, on a hill.

[1255] I don't know exactly how I'm sure how it translates.

[1256] I tried reading a couple of times.

[1257] I still don't get it, but everything else I fully enjoyed.

[1258] And like for one of my birthdays as a kid, I got like their entire collection, like occupied a giant shelf in my room.

[1259] And then over the holidays, I just like, you know, my parents couldn't drag me out of the room and I read the whole thing cover to cover.

[1260] And it, it, I really enjoyed it.

[1261] And that's one more.

[1262] For the third one, you know, maybe a little bit darker.

[1263] But, you know, it comes to mind is Orwell's 1984.

[1264] And, you know, you asked what made an impression on me and books that people should read that one, I think, falls in.

[1265] the category of both, you know, definitely as one of those books that, you know, read and you just kind of, you know, put it down and you're in space for a while.

[1266] Yeah, you know, that, that kind of work.

[1267] I think there's, you know, lessons there.

[1268] People should not ignore.

[1269] And, you know, nowadays, with, like, everything that's happening in the world, I, like, can't help it, but, you know, have my mind jump to some, you know, parallels with what Orwell described.

[1270] And, like, There's this whole concept of double -think and ignoring logic and, you know, holding completely contradictory opinions in your mind and not have that not bother you and, you know, stick into the party line at all costs.

[1271] Like, you know, there's something there.

[1272] If anything 2020 has taught me, and I'm a huge fan of Animal Farm, which is a kind of friendly, as a friend of 1984 by Orwell, it's kind of another thought experiment of how our society, may go in directions that we wouldn't like it to go.

[1273] But if anything that's been kind of heartbreaking to an optimist about 2020 is that that society is kind of fragile.

[1274] Like we have this, this is a special little experiment we have going on and not, it's not unbreakable.

[1275] Like we should be careful to like preserve whatever special thing we have going on.

[1276] I mean, I think 1984 in these books, The Brave New World, they're helpful in thinking, like, stuff can go wrong in non -obvious ways.

[1277] And it's like, it's up to us to preserve it.

[1278] And it's like, it's a responsibility.

[1279] It's been weighing heavy on me because, like, for some reason, like, more than my mom follows me on Twitter.

[1280] and I feel like I have like now somehow a responsibility to do this world.

[1281] And it dawned on me that like me and millions of others are like the little ants that maintain this little colony.

[1282] Right.

[1283] So we have a responsibility not to be, I don't know what the right analogy is, but put a flame thrower to the place.

[1284] We want to not do that.

[1285] And there's interesting, complicated ways of doing that as 19.

[1286] 1984 shows.

[1287] It could be through bureaucracy, it could be through incompetence, it could be through misinformation, it could be through division and toxicity.

[1288] I'm a huge believer in that love will be somehow the solution.

[1289] So, love and robots.

[1290] Love and robots, yeah.

[1291] I think you're exactly right.

[1292] Unfortunately, I think it's less of a flame thrower type of an extra.

[1293] I think it's more of, in many cases, can be more of a slow boil, and that's the danger.

[1294] let me ask it's a fun thing to make a world -class roboticist engineer and leader uncomfortable with a ridiculous question about life what is the meaning of life demetri from a robotics and a human perspective you only have a couple minutes or one minute to answer so i don't know if that makes it more difficult or easier actually you know they're tempted to quote one of the stories by Isaac Asimov, actually, actually titled, appropriately titled, the last question, a short story, where, you know, the plot is that, you know, humans build this supercomputer, you know, this, this, this AI intelligence.

[1295] And, you know, once it gets powerful enough, they pose this question to it, you know, how can the entropy in the universe be reduced?

[1296] Right.

[1297] So, you know, computer.

[1298] replies, as of yet, insufficient information to give a meaningful answer, right?

[1299] And then, you know, thousands of years go by and they keep posing the same question.

[1300] The computer, you know, gets more and more powerful and keeps giving the same answer, you know, as of yet insufficient information to give a meaningful answer, or something along those lines, right?

[1301] And then, you know, it keeps, you know, happening and happening, you fast forward, like, millions of years into the future and, you know, billions of years.

[1302] And, like, at some point, it's just the only entity in the universe.

[1303] It's, like, absorbed all humanity and all knowledge in the universe, and it keeps posing the same question to itself.

[1304] And, you know, finally it gets to the point where it is able to answer that question.

[1305] But, of course, at that point, you know, there's, you know, the heat, death of the universe has occurred, and that's the only entity, and there's nobody else to provide that answer to.

[1306] So the only thing it can do is to, you know, answer it by demonstration.

[1307] So, like, you know, recreates the big bang, right?

[1308] And resets the clock, right?

[1309] I can try to give kind of a different version of the answer.

[1310] You know, maybe not on the behalf of all humanity.

[1311] I think that might be a little presumptuous for me to speak about the meaning of life on the behalf of all humans, but at least, you know, personally, it changes, right?

[1312] I think if you think about kind of what gives, you know, you and your life meaning and purpose and kind of what drives you.

[1313] it seems to change over time, right?

[1314] And the lifespan of, you know, kind of your existence.

[1315] You know, when you just enter this world, right, it's all about kind of new experiences, right?

[1316] You get like new smells, new sounds, new emotions, right?

[1317] And like, that's what's driving you, right?

[1318] You're experiencing new amazing things, right?

[1319] And that's magical, right?

[1320] That's pretty, pretty awesome, right?

[1321] That gives you kind of meaning.

[1322] Then you get a little bit older, you start more intentionally learning about things, right?

[1323] I guess actually before you start intentionally learning, probably fun.

[1324] Fun is a thing that gives you kind of meaning and purpose and the thing you optimize for, right?

[1325] And fun is good.

[1326] Then you get, you know, start learning.

[1327] And I guess that this joy of comprehension and discovery is another thing that gives you meaning and purpose and drives you, right?

[1328] Then, you know, you learn enough stuff and it, you want to give some of it back, right?

[1329] And so impact and contributions back to, you know, technology or society, you know, people, you know, local or more globally, you know, becomes a new thing that, you know, drives a lot of kind of your behavior and, you know, something that gives you purpose and that you derive, you know, positive feedback from, right?

[1330] You know, then you go and so on and so forth.

[1331] You go through various stages of life.

[1332] If you have kids, like that definitely changes your perspective on things.

[1333] You know, I have three that definitely flips some bits in your head in terms of kind of what you care about and what you optimize for and, you know, what matters, what doesn't matter, right?

[1334] So, you know, and so on and so forth, right?

[1335] And it seems to me that, you know, it's all of those things.

[1336] And as you go through life, you know, you want these to be additive, right?

[1337] New experiences, fun, learning impact.

[1338] Like, you want to, you know, be accumulating.

[1339] I don't want to, you know, stop having fun or, you know, experiencing new things.

[1340] And I think it's important that, you know, just kind of becomes additive as opposed to a replacement or subtraction.

[1341] But, you know, those few is probably as far as I got.

[1342] But, you know, ask me in a few years, I might have one or two more to add to the list.

[1343] And before you know it, time is up just like it is for this conversation, but hopefully it was a fun ride.

[1344] It was a huge honor, Dmitja, as you know, I've been a fan of yours and a fan of Google Self -Diving Car and Waymo for a long time.

[1345] I can't wait.

[1346] I mean, it's one of the most exciting.

[1347] If we look back in the 21st century, I truly believe it would be one of the most exciting things.

[1348] We descendants of apes have created on this earth, so I'm a huge fan, and I can't wait to see what you do.

[1349] next.

[1350] Thanks so much for talking today.

[1351] Thanks.

[1352] Thanks for having me and it's also a huge fan doing work honest and I really enjoy it is.

[1353] Thank you.

[1354] Thanks for listening to this conversation of Dmitri Dolgov and thank you to our sponsors.

[1355] Trial Labs, a company that helps businesses apply machine learning to solve real world problems, Blinkist, an app I use for reading through summaries of books, Better Help, Online Therapy with a licensed professional and cash app.

[1356] The I use to send money to friends.

[1357] Please check out these sponsors in the description to get a discount and to support this podcast.

[1358] If you enjoy this thing, subscribe on YouTube, review it with $5 ,000 on a podcast, follow on Spotify, support on Patreon, or connect with me on Twitter at Lex Friedman.

[1359] And now let me leave you with some words from Isaac Asimov.

[1360] Science can amuse and fascinate us all, but it is engineering that changes the world.

[1361] Thank you for listening and hope to see you next time.