The future of driving is autonomous with Dmitri Dolgov of Waymo

Hello and welcome to how I built this lab. I'm Guy Raz. There's no driver. How crazy is this? Crazy. This is your first time ever. You're gonna look at this as an adult when all cars are driverless. You're gonna be like, yep. I did this driverless drive with Daddy.

So it's probably hard to tell from the sound alone, but what you're hearing is audio from a ride I recently took in a Waymo Autonomous Taxi. My son and I were in San Francisco heading across the city. These taxis are now all over San Francisco and Phoenix, and Yes, there is no human driver. There's no human in the car at all, unless you count me and my son, who were passengers. It's a ghost driver, and if I'm being totally honest... They're amazing. The Waymo Taxi is now available 24-7 in San Francisco and Phoenix, and it operates similar to Uber or Lyft. Pull out your phone, you fire up the app, order the car, and it magically appears within minutes. Except

It's fully autonomous. Waymo was spun out of Google a few years ago, but the project to Build autonomous vehicles at Google actually goes back to 2009. One of the engineers on that project was Dmitry Dolgov. He's been with Waymo from the very beginning and today He's the company's co CEO.

Dmitry was born in Russia a child of two physics professors. He went to high school in the US and eventually returned to Russia to attend the Moscow Institute of Physics and Technology. Dimitri then came back to the US to get his PhD. In 2006, he took part in a competition called the Urban Challenge, which was sponsored by DARPA, the US defense agency. The idea was to build an autonomous vehicle that could make it through a series of obstacle courses.

And the grand challenge, the the grand challenge was to drive 150 miles through a desert. completely aesthetic environment, but you know, until then robots were not capable of doing that. So it took two tries in 2004 and 2005 to accomplish that. So the next step was let's make it a little bit more interesting. Let's make the environment dynamic. All right. Let's add you know some rules of the road, stop signs, other vehicles. But do it in a controlled Yeah. uh in that mock city. So that that was the challenge and we build a car to do exactly that. We equipped it with a bunch of uh sensors, you know, lighters, lasers, radars, cameras, you know, a computer and then road software. that would allow it to follow the rule understand and follow the rules of the road and interact with other dynamic actors, whether you know human or other robots.

Um I I'm looking at one of the vehicles that you were that you helped. put together it's called the the junior. And uh this of Volkswagen w uh Passat, I think. And um

I think came in in second place. Um this is like two thousand Seven. And and was was the the the underlying technology Basically what we're talking about today was it

Pretty similar, or was it like a crude version? Uh you know, yes and no. Depends on how you talk about like the big A car. Plus sensors. Plus computers.

Plus software. Yeah. That stays. Right. Uh But of course, you know, it's what it almost eighteen years and then everything has changed, right? The the sensing technology has

uh gone a very, very long way. Uh computers, you know, have evolved software, right? All the breakthroughs especially in A. I so you know that that led to uh many, many breakthroughs in the area of software. So now the system we have kinda has the same components. But yeah, all of the components and how they work together is You know, qualitatively drastically different. Yeah, it's amazing how many

sort of self driving car companies they that they spawned. I mean, we we've talked to Kyle Vogt of Cruz, he was in he competed in the two thousand four and two thousand five challenges. Dave Ferguson. of Nuro, which is another um company was part of the Carnegie Mellon team, Chris Earms, who also is uh obviously worked with you at Waymo, he was he's now with Aurora, um, was was also a part of the Carnegie Mellon team. It's amazing how many

Like you knew all of these people. Um, and this is as you say is a small community of people working on this huge challenge. um to basically create vehicles that could be fully autonomous. That's right. That's right. And of course, uh worked very closely with for many years with you know Dave and Chris and you know they're they're great friends. So yeah, it's a a a lot of innovation and a lot of companies and In progress.

Yeah. came out of those early days of the Arco Gen Grand Challenges. You and um Sebastian Thrun and Chris Ermson, you all went to go work for Google. Um, and um and you were part of that founding team, which was uh uh at the time a secret project. Google's self driving car project and

At that time it's uh amazing to think two thousand nine,'cause it s it seems like ancient history now. You were given a charge by Larry Page and Sergey Brenn, the co founders of Google. Um you had two years. To

Uh w you two years to essentially accomplish two things. W do you remember what they were? Yeah, I I remember them very well. Uh the first one was to drive a hundred thousand miles in uh autonomous mode. That was way more than orders of management more than you know what anybody has done uh at that time. And the second one, and you know, actually that one turned out to be more interesting and much more challenging was to drive Ten routes. Each one was about a hundred miles long.

And Yeah, very carefully selected. I think but you know by Larry and Sergey. uh personally and they're you know fairly devious in uh how they

Yeah, created them uh to make it interesting. Um and the goal was to drive each one from beginning. To end. In full. autonomy with no human intervention. Which at that time, yes, seemed

you know, almost impossible. I remember we had a lot of people, you know, even experts in the field kind of you know laugh at us when we attempted it. You two years to accomplish this. I think you guys finished it. with three months left. Like you actually I mean, going into this project, do you remember thinking, we're never gonna do this in two years? I was actually Yeah, maybe naively, but fairly optimistic.

Right. And It was hard. And they're all different. And you know, yes, there are many moments where we would, you know, finish one route and we

would think ahead of like what would it take to do the next one. And Yeah, we would. Write some software. Yeah, we would you know collect some data.

Test it. And then would actually try it and have Then we would yeah, hit a bunch of Oh crap moments of like okay, wow, this is way more difficult than we expected in the even those early days. So so when you did finally accomplish it W did you g were you guys able to have a party and celebrate or was it just like nice job, now we gotta we gotta keep it quiet? Uh.

Yes, both. We we celebrated. And you know, I felt like it was a big accomplishment. And but yeah, we kept it quiet. Uh but y y you're like that's what made it So Incredible. fun. I guess you know that that phase was

Yeah, one of my Favorite. Yeah, phases of the whole project. It's like The early days. of a startup you're up against.

But might seem like an impossible goal, but it's very clearly defined, right? You have a very clear milestone, a very clear uh goal you are singularly focused on it. You have a small team. You know, everybody's working around the clock twenty four seven.

And you know, sprinting. together and every day and every hour you are like you're prototyping, right? You're learning. Uh so you can move incredibly fast and yeah, like every day of every hour you're making amazing progress and you're learning new things. So that was that was uh that was a total that was a blast. And every day probably presented a new series of challenges. Do you remember What was one of the hardest challenges that you had to figure out? Like something that just took longer than you thought, that you just

It just didn't It wasn't coming together quickly. on um working on I'm trying to make this happened. Do you do you remember something you worked on that you just it was like such a hard problem to solve.

Where's the one? I mean we ha we had a we had a number. Some were more fundamentally challenging, some were kind of Even Comical. Subbacks.

Where you know you would do a I remember one one w one of the routes was driving on all of the freeways and crossing all the bridges in the Bay Area and you know it's about a hundred miles and you go through all of the challenges and you were we're tempting this this drive and yet the car is Doing a good job. Yeah, it's ha handling merges and you were you or somebody else was sitting in the driver's in behind the wheel just in case. That's right. That's exactly right. That's exactly. And that was kind of what might might made those early days a lot of fun, is that you would

If you do everything, right, you would be, you know, uh putting some hardware in the car, you would be calibrating the hardware, then you know, the next hour you're writing some software, you know, whether it's tools or something for the car to actually make decisions. And then you know, you get in the car and yeah, you you you you give it a try. So anyways, on this route, we are driving along, yeah, it's doing a job. We get almost to the very end. And so at this point you are gonna you know holding your breath. You're we're waiting. For yeah, the last you know, mile or so off that you know, a hundred mile. And the way that particular run Uh

Uh supposed to finish is that we're coming. Down the Golden Gate Bridge. Yeah, uh into the city. And there are a set of Tallbooth.

At the yeah, when you're as you as you go into the city. That's right. at the end of the Golden Gate Bridge and they're They're narrow,'cause I go use them almost every day. That's exactly right. And Yeah, the one

that our car wanted to go through was closed. At the time. And it's just not something that We ever encountered and we thought about and Oh right,'cause it didn't it couldn't figure out like the X, the red X or the the air the green arrow like it didn't

You couldn't recognize w what those meant. That's right. Nor is actually a gate, so you know what it would stop, but it feels like it would not change lanes and pick a different one. Like, oh yeah, my God. Yeah, now back to square one. Because probably every uh you know, we this impossible to imagine every eventuality. But something's gonna come up. That's exactly. That's exactly right. And yeah, we would have to deal with you know high speed traffic on freeways. Yeah, we had one route that went Yeah, kinda took this windy road, uh, from you know the Bay Area to

uh highway one, you know, through the mountains and you know, we were driving along and then a bicycle. fell off the truck in front of us. So Yeah, that's not something that that time the car could deal with. We had another route that went through downtown San Francisco, uh and uh you know famous Lombard Street. That's very narrow, very windy, and has you know some of the most adventurous, you know, pedestrians and tourists uh in the world. So challenges like that. Like right. So that that's what made it so challenging and so interesting that it was the kind of the breadth of the the experience. And probably every single ride posed a new series of challenges. Like I I like, for example, you know, some lanes are shoulder lanes, but then during

um heavy traffic that are open for driving But the lines are painted on them in such a way that it doesn't seem like a lane. Like a human could figure that out. But an autonomous vehicle probably at that time was like wait. This doesn't make sense. The the lines

Don't make sense. They don't align and maybe the car was confused. That's right. That's right. And that would be Yeah, not the kind of situation.

Or condition that at that time we would Yeah, we're able to solve robustly. I was you know the goal there was to learn and do like we had to do the route once, right? So if you failed at something, you would go, you know, improve the system and you would try it again. So there was uh you know, uh a very well scoped milestone because you it A hundred miles.

Yeah, is nothing. If you want to build a production system. Right. Uh but it in those early days it was long enough that it actually forces you to very deeply think and tackle some of the most fundamental, most important challenges that exist. We're gonna take a quick break, but when we come back, more from Dmitry Dolgov on how the engineers at Waymo decided to go after the big win and create a fully autonomous vehicle. Stay with us, I'm Guy Raz, and you're listening to How I Built This Lab.

Hey guy Ros and HIBT listeners, my name is Richard Crowdy and my favorite episode as a business professor is the founding of BET because it summarizes business schools perfectly in keeping revenues up and keeping costs down. It also shows that a complex company like a media company can be founded by an individual that bootstraps it with the proper type of innovation. Bob Johnson is also a legend and he's hilarious. I've actually been listening to my top fifteen to try to come up with a list for my business students. So thank you, guy, Roz and cheers. If you want to share your favorite episode of how I built this, Record a short voice memo on your phone telling us your name, where you're from, what your favorite episode is, and why.

A lot like the voice memo you just heard. and email it to us at hibt at id.wonder.com, and we'll share your favorites right here in the ad breaks in future episodes. And thanks so much. We love you guys. You're the best. And now Back to the show.

Welcome back to How I Built This Lab. I'm Guy Raz. My guest today is Dmitry Dolgov, the co-CEO of the autonomous vehicle company, Waymo. I guess in those early days the the Google strategy was really on creating driver assist technologies, not necessarily fully autonomous cars, but cars that could make it easier for drivers to navigate. Basically what you've got in you know in like a Tesla now or some other cars that have driver assist technologies. Um but I guess around twenty thirteen, um there was a conscious decision to pivot the focus because most of your testing was done on freeways. That's right. But what I think m you guys pr and most people involved in the space realize is that

complexity. You've got to test these in cities. They've got to navigate city centers and stoplights and left turns and intersections And so what happened in two thousand thirteen to kinda get you guys to focus to shift your focus on building a fully autonomous vehicle instead of just

driver assist technology. Right. Uh I would say there are yeah a number of things. As you mentioned, the first phase was, you know, just learning and understanding the complexity of the problem. Then we said, Let's try to build a product and uh was the driver uh system. You know, we tried that. That was actually reasonably successful. We had ran a pilot where we gave these cars to about you know a hundred Google employees and they could, you know, use them to commute and Um take them around on on their daily trips. Um So

That was Yeah, around two thousand eleven, two thousand twelve. So then and you know two thousand thirteen we looked at the whole thing and we made this decision to go after full autonomy. So the reasons were yeah several. One was What we learned. from people using the driver as the system. They would overtrust it. They would you know would

Yeah. put on makeup and, you know, text and you know, one guy actually fell asleep. Uh, you know, the car handled everything fine, but that does, you know, not what we wanted to see. So that wasn't one of the reasons. Uh another one was that you know we actually made Progress on the Core. you know, most difficult aspects of this driving task um in on surface streets. So that you know gave us uh a bit more optimism of you know going after like the the big prize and this also at the same time the field was moving forward. So if we wanted to really make an impact um in make

Yeah, improved transportation. you know, globally, I thought, okay, you know, the the field, you know, driver systems, you know, we'll develop there, you know, other companies will be you know pushing in that direction. But let's play to our strength. Right. And let's go after the kind of uncompromising, you know, big win uh and you know big unlock in in the space, which is full autonomy and actually building

Yeah, what we now call the Wainwood driver that's responsible for the whole task of driving beginning and to end with you know no human behind the wheel. What do what do you remember about the conversations and the ambitions? Was it like Hey. You know, let's build the technology for fully autonomous vehicles.

that we can then maybe one day license or was it no Let's Basically. become a car company. Let's like make fully autonomous technology and purpose built vehicles.

And Uh, you know, and eventually sell these to consumers that they can have for themselves. Discussed? I don't think we ever

Seriously. Entertained building a car. Okay. Just never made sense, you know.

to me, you know, or others to you know do that. Like we're not a car company. Building cars hard. Yeah, there's many companies that you know spend a hundred years getting very, very good at it. You know, we did in the early days. of design and um uh A low speed vehicle, you know, that we called the Firefly because we needed to take that s you know first. That was our zero to one moment of full time. It was a prototype. It was like that you showed uh you sho we publicly should display the in twenty fourteen. But that wasn't the goal. It wasn't we're not gonna build cars. But we're gonna build the technology that can power any car to become autonomous. That's exactly right. That's exactly right. And that's you know, uh it was the evolution of the thinking then then then.

Crystallized. in you know the Wimu. mission which is to build. The way more. Driver.

And uh Yeah over time. deploy it in different products and you know different commercial applications. Yeah, whether Right healing. trucking deliveries and eventually, you know, personally owned vehicles. And that that's the that's the path that we

Yeah. set for ourselves and the path that we've been on since uh about that pivot in twenty thirteen. Now. in some ways the real work begins because

between twenty thirteen and let's say twenty twenty three. Right, which we're gonna get to'cause it's Incredibly exciting. what is going on now in San Francisco and Phoenix and a you know a couple of other places. But

That tenure period was It uh probably all of a sudden you're back to kind of startup mode because You've gotta get these cars, this technology. To be

Absolutely foolproof perfect. But in complex Environments like San Francisco. Um, which I think is uh next to New York is one of the most complex driving environments. in the United States.

Tell me a little bit about the process that then began. Was w is that what happened? Was it a shift to like, okay Let's see what these can do in cities. Uh that's exactly right. Uh yeah, we didn't start with the full complexity of San Francisco, uh, in those early days. So we can kind of think of you know that ten years as maybe three phases and kind of that correspond to three generations of our technology. Three uh generations of the you know uh of the of our driver. You know, that first one was on that low speed vehicle.

Yeah, the the f the firefly. That's what we call the third generation uh of our driver. And by the way, just to clarify, the Firefly probably was designed to be like a campus type of vehicle, right? Like in a university or office park. Not necessarily in the city. That's exactly right. I mean it was a low speed vehicle and you can only uh move up to twenty five miles an hour. So we're thinking, you know, maybe large retirement communities or campuses. That's exactly that would just you would get off and on. That's exactly right. And in that first phase on the third generation of our driver, the goal

was to actually build something. that can take, you know, a fully autonomous trip. I w and we actually did that in twenty fifteen. Uh and yeah, we put a Uh a friend of our project, uh his name was is Steve Mann, happens to be blind, and in twenty fifteen he took the first autonomous ride in Austin, Texas. Uh the first person, the first public That's right.

Um, I remember this. This is in uh'cause this is a YouTube video of it. And uh and yeah, it was in Austin and um Yeah, I mean That was kind of a big deal. Uh yeah, it was so it was a huge moment when we're able to do this first ride.

in twenty fifteen was uh yeah uh uh a a big celebration after that. Yeah, I can imagine. Um, so a at that point you in your team and built this this third generation version of the Waymo driver that could actually Take a fully autonomous trip, which uh is is a big deal. Um and you were essentially betting that this

vehicle would keep passengers safe on a real on real city streets, right? Like what what went into the safety design? That's right. That's right. Uh it was also That early.

Kind of. ps forced us to start thinking about this. Very Fundamental question.

I go what does it mean for a self driving fully autonomous vehicle to be ready? I command set out on this path in twenty thirteen. So okay, you know, we're gonna Yeah, we're gonna need a car. It needs to have, you know, a bunch of safety systems. Does that exist? No. Okay, well, let's design one and work, you know, with partners to manufacture it. So we put a lot of thought and work uh into making it safe. It had like a foam. Yeah, front, it had a plexiglass window. The sensor podes you know were attached with magnets.

So they you know could uh detach if something were to happen. Uh and then we Uh the sensors. at that time and didn't have the level level of reliability or capability that we would need that we would trust, you know, to go to full autonomy. So we build, you know, uh if that generation had our own lighters and you know, um custom uh sensor suite. And then of course the software that we had to build. Like that that none of that existed. So let's let's talk about the technology for a moment.

Um because um you mentioned lidar and radar and and some of these things we we know what they are. LIDAR, for example, uses basically light lasers to measure distance. Um uh and you've got radar technology Cameras. um all over the car. Can you just kinda break down

how they work. I mean a a lot of people who drive Teslas, for example, I have one they uh if we if you use the f what what what Tesla calls full self driving uh which is not really full self driving, but um They rely uh primarily on cameras. They don't use LIDAR or radar technology. Tesla argues that that is

The other ones are just r redundancies that are unnecessary. Tell me w how your technology works and and why you think it's better. Well these uh censorship. Or Okay.

uh fundamentally different and complementary physical. Properties. I mean uh yeah, so cameras Uh give you Yeah.

high resolution and you know the richness. Of colour. And how many cameras, by the way, on a Waymo vehicle now? On the current generation we have kinda all of them, including the internal ones we have twenty nine. Uh cameras. Twenty nine cameras. Okay. Yeah. Yeah, uh but they're passive.

And so right next you know, somebody else has to bring the light, whether it's your headlights or the sun. Radars and lighters. in contrast are active sensors. So yeah, they blast you know their own energy uh on the wall and then they you know, get returns and from that they can make sense of the environment. And they use you know different wavelengths. So they I can kind of punch through fog or rain much better than a light or a camera. I think as a you know a human driver.

Sure, you can you know how difficult it may drive to to drive at night, right? For example, at night if you have, you know, uh somebody in an oncoming car with their headlights on. you know, high beams. It kind of blinds you. It's very hard to see. Uh it doesn't affect, you know, uh radar. Uh or lighter or similarly, you know, driving in you know dense rain or fog. So this is why you know we think kind of using all of the sensors and Yeah, fusing them in our AI and ML models. Yeah, so that you can kind of extract the best signal and see the world.

in the best possible way uh gives us an advantage. And yeah, really you can build, you know, a prototype or can you know build a driver's system, you know, without needing all of that. Uh yeah, extra. capability, extra redundancy, but if you really want to take the driver out and go for full autonomy, yeah, it that gives you a boost. Alright, so so just to to to clarify, like you ba basically you have the first successful ride with a third generation vehicle in twenty fifteen.

And then in twenty sixteen you way much just spins out of Google as a as an independent company. And then The next year you become the first company to start like regularly operating these these A Vs In Chandler, Arizona. These are m even more advanced than the previous version.

Wha was was the idea of like, hey, let's get these in really good shape and then, you know, we'll we'll turn them into like Like Ubers. That's right. That's right. Uh that point we were

And pretty clear that that was going to be our first deployment. our first product. And I essentially that's what we launched in Chandler. Yeah, that that's when we can create a the Wayma one. product and the the Wayma One uh application. And in twenty eighteen in Chandler we started uh offering the service. to external writers.

Uh and then in 2020. In Chandler, Arizona, we launched, you know, the first fully autonomous uh right healing service that was open to the public. Anybody could just download the app and call a car. Uh that was our fourth generation, you know, Pacifica mini van and the empty car would go show up and take anywhere. So then we made this Yeah, decision. That

Yeah, it was. not the best path forward to kind of incrementally grow. and scale that system. made the decision to make a hop to what we call the fifth generation of our driver. That's on the J L R uh ey pieces and Pacificas and a whole lot of the the Jaguar. That's right. Uh and you know we said hey yeah it's

Let's take a big step. gonna be you know a different car. It's gonna be a new generation of hardware. It's gonna be very different software with you know big bets on in uh state of the art AI and Okay, let's go after the full complexity of the problem, including the full density of downtown Phoenix and downtown San Francisco. Uh so that's what we were working on, you know, on that time frame to then on that new generation or the Whever to launch

Uh the Waymo one. Uh service. We're gonna take a quick break, but when we come back, how Waymo One works today and Dimitri's take on the future of autonomous vehicles. Stay with us, I'm Guy Raz, and you're listening to How I Built This Lab. Welcome back to How I Built This Lab. I'm Guy Raz. My guest is Dmitry Dolgov, the co-CEO of the autonomous vehicle company Waymo, which has started to roll out Waymo One, its autonomous ride hailing service in San Francisco.

the city gave you permission to to be a ride hailing service for twenty four hours a day. Uh and I get I've I've used them probably a dozen times now. Um their Jaguar S C Bs All electric. Um and they driving around the city and you just order it.

Like an Uber. You just go out and you you know, go on your app and it comes and then you unlock the door with your And uh you hit start and it goes. And I I can I can I

can't even tell you how many times i the first question I get from people is aren't you scared? Aren't you terrified? When you get in one of these things Um and so I have my answer.

So what's your answer when people say that to you? Like this just seems terrifying for a machine? To drive you around a city like It seems so scary like He could you know, he could just go haywire and and and drive off a bridge or something. What do you w what do you say to people when they ask you that?

Oh you know, I I at this point I think I'm much more anxious about human drivers than I am about The Waymo driver. So I have you know full confidence, but of course, you know, I've been uh in those cars.

Yeah. Many, many times. Yeah, hundreds, thousands of times over the years. But what we see with other people, like it's very natural to have that anxiety. It's a very different, very new system and new product. Um, but we what we very consistently see, and I I guess I would Love if you're not going to be able to To hear if that matches your experience. But once people get in the car.

Just you know, after a couple of minutes. They get very comfortable. And they go back to, you know, doing whatever they want to be doing, like in the back of the phone. Yeah, exactly. I know. I know. I w I took my son in it, um And drove across San Francisco and

Yeah, within like thirty seconds or forty five seconds. The first time ever he's in there, he's like back looking at his phone. And it's true. It it's you get in it. And then it goes. And it's really cool you're looking out and looking at the steering wheel turn for the first forty five seconds and then you're done. And then you're just like going back to what you were doing, answering emails or whatever. But it is it is amazing.

But it you know. How sort of bizarrely ordinary it feels. After you after a couple of seconds. Uh Yeah, but that's great.

Yeah. Uh I mean that that's uh that that's incredibly exciting and I think some people kind of uh draw this parallel that you kind of your brain switches to Passenger mode. Yeah. And yeah, you do get all of these.

benefits of, you know, privacy, whether you want to have a conversation with somebody in the car the that you're in the car with or you want to make a, you know, phone call. Um, and I mean yeah, that's exactly Yeah. Why we're why we're so excited about this project. Alright, so Dimitri, in general Right.

I I'm I'm optimistic when it comes to technology, but I have to also admit that I healthy skepticism is important, right? I mean, we've been promised technologies that were gonna make our lives better and change the world only to be Yeah. you know, sorely disappointed by them. I think it's really amazing what's what's happening with

Autonomous vehicles. But I I I guess I wonder You know Do you understand some of the cause? for concern or some of the skepticism around it.

Uh yes, I think it's very natural. That is very important. New technology. Uh it's Yeah, very different.

So I think If you look back in history It's Very

common that when a new thing comes around Uh There's skepticism. Yeah. There is, you know, excitement, but there's also, you know, a lot of sensitivity.

But I wonder, I mean You know, it's one thing to to say, Hey, you know, don't worry, this is gonna be okay. But how can you guarantee or convince people that the technology couldn't be misused. Couldn't be

You know. Manipulated in a way that endangers human life. Well I guess yeah the we should start I would start with the status quo. Right. We are not okay.

Right. Uh if you look Yeah. Just How many Uh lives are lost.

to the transportation system that we have today. I'm sure you've heard these numbers before, but you know, well known. that in the US alone, more than forty thousand people die every year. I just need to take a step back. This is It's kind of insane.

Yeah. If we were to invent Cars. today our transportation system like No way we would Allow this.

Right. And I think we over time, which is kind of slowly boiled ourselves, society, to Yeah, uh accept that. Right. I think as a first

order impact of this technology. Yeah, we can we can do better. We can do much better. And we are, you know, we are seeing that today. Uh we have driven well uh north of five million fully autonomous miles today. And we've shared some data from uh the safety impact that our cars have. And I think at this point we have a fairly robust body of evidence that shows that, you know, our cars

actually have very clear safety benefits where they operate. How do we think about and there's no easy answer right now'cause it's all both an ethical and a legal question, but How do we think about liability, right? I mean, if let's say Uh I own a car that is fulonomous with Waymo technology.

Let's say in ten years from now. And I'm in the back and I'm just doing my work. Or I'm asleep. you know, which I I should rather be doing. I'd I I'd just go to sleep and let it drive me from San Francisco to LA, which would be great. But it gets in an accident, uh just a very a fluke accident. Maybe another car hits it or something.

Um, how do we account for Liability. I mean who's responsible? Is the owner of the car? Is Waymo's technology? Like, how is that gonna work? for the actions of you know the W driver. Yeah. the responsibility Yeah, lies.

Yeah, with way more. Uh if it was Yeah, the fault. of yeah another actor. another driver, then you kind of follow the established

Yeah, processes. uh that are you know well understood and well studied by for example insurance companies that that's you know they have you know decades of experience of kind of uh evaluating exactly that you know that qu the question and this is where I think that uh study. that Swiss re has done was very encouraging that when they looked at you have almost four million miles

Of Yeah, our fully autonomous operation, they found that massive reduction in, you know, 100% reduction in the bodily injury claims and you know fourx reduction and property damage, right? So that then you can apply like you can marry the two and you're starting to see the benefits.

So if Waymo is I mean basically If if in in a future scenario, somebody's in a Waymo car, the Waymo vehicle crash and it's There's a some kind of fluke. and Waymo is responsible. I mean you guys have to accept that liability, but I guess

in order to get to that position, you have to be Rock solid confident. That that will never happen. Well, that's what we must spend. That's one of the hardest

questions that we spend, you know, more than a decade working on. We have we've developed uh A very robust multifaceted uh readiness and safety uh framework. And actually that's something we shared publicly. And yes that what we see in all of those methodologies as we improve and validate our system, that at the end of the day is what gives us confidence in the performance of the system. How many cars do you have on the streets of San Francisco now?

Uh we have a fleet in San Francisco of about two hundred fifty vehicles. You know, they're not all out uh at the same time, but it gives you an approximate uh order of magnitude. And Phoenix. About the same. Uh wanna say about a couple hundred cars of all. So he here's a question. I mean is is the part of the business model right now for Waymo to become like

a ride hailing service, like You know, I'm sure I know you've got a partnership that you announced with Uber, but Um, I mean i it is the idea that, you know, in ten years time Uh this is gonna be the primary Your primary business or or a part of what you do or

Give me a sense of of like are is Waymo going to be a right hailing service or is it going to be a technology company that licenses its technology to both ride hailing services and Auto you know, automobile manufacturers and and others. Oh.

We Think of ourselves as Yeah. Building a generalizable WAMO driver. And the business model is to deploy it in different

Product lines different commercial applications. There's three main ones. Uh right healing. That's what we call way more. One. Yeah, trucking and deliveries, moving goods. And then the third one is

personally owned vehicles. So that's the long term vision. Like we want you know, the driver there's trillions of miles uh being you know travelled. I think almost three m trillion uh in the US alone. You know, much more in the across the world. So we want to have you know, a positive impact.

On Yeah, some meaningful fraction of all of those miles. So you know, the first business and the first product is Wing Wan. And you know, right hailing. But uh we're exploring, you know uh different uh you know other different partnerships. Uh the Uber partnership that we just launched and where it is. And and that partnership, by the way, is w what is that gonna look like? It's gonna be the Uber app will also hail at Waymo Cars?

That's right. That's right. you can use the Uber app and uh get a fully autonomous vehicle. This is not

um uh a a money making operation for for Waymo right now. I mean the the right hailing service is still in its infancy, but Um, you know, there's been billions of dollars invested in in into Waymo. Um and I have to imagine ride hailing is not where you're gonna make your money. It's gonna be from selling this technology. Um so what tell me just from the business perspective

When do you see a Path to profitability. Well you know, a challenge a little bit that ringing right hailing is a massive opportunity. Yeah. It is a very big market today. Um, but yeah, yeah, th it's it it's growing. There's expectations that it's gonna be significantly bigger by, you know, the end of the decade. Uh, but if on top of that, if you factor in, you know, the

Yeah, the benefits and the you know. positive economics that fully autonomous vehicles can bring to the table, it and there's potential for that to, you know, expand. Quite a bit. So yeah, we were very, you know, laser focused uh on that as our primary Uh business line.

Yeah, beyond that, yeah, we want to uh pursue trucking and deliveries and then eventually personally owned vehicles. But right healing like was I would not. Dismiss that at all. So probably unrealistic to say that within ten years

ordinary people could buy. a fully autonomous vehicle for themselves, but probably not unrealistic to say that in ten years from now, in most major urban centers in the US. There will be autonomous taxis available for anybody to use. I definitely agree with the latter.

And I would not dismiss the former. Ten years is a you know reasonably long time and you know things uh can happen nonlinearly. So uh final question for you. Um so I mean if the future is autonomous. And I think it is. I think I really do, I'm convinced. I think anybody who uses One of these taxis.

Well see it. It's it's so clear. at at least to me, you know, you go in it and it It's a clean car. It's a very good driver. It's a defensive driver, but it's also Not overly defensive, so it's not timid. Um, it's like a very good uh taxi driver.

A better. So that I think is is the future. So Dimitri, I I know that, you know, there are plenty of people who love driving, right? Who love The experience of controlling their car and

Th th those people will continue to want to have that ability and they will. But in your view. Is are we looking at a future where most people are gonna be driven by their cars. In the long term future, I think, yes. I think that that's where

Yeah, we're we're heading. Maybe you know, taking your car, uh in the future to a racetrack and driving it manually. That's gonna be the And and a and an exciting thing.

rather than the mundane kind of boring task of driving commuting. And if you uh imagine a future where You know, a large Um the fraction. of your cars in the road. Are

you know, fully autonomous or at least smart enough, then you can start doing things where you're optimizing, you know, more globally. I they can coordinate, you know, their speed. You can you know connect them to smarter infrastructure and actually overall increase the throughput of your roads and could, you know, kind of increase the throughput of your transportation system. And you know, uh farther out in the future if you look uh There today. personally owned vehicles.

um are can sit around for ninety percent of their you know their lifetime, right? You know you take'em you know home or work and then you park it and then yeah nine out of ten hours is just sitting there. So if that changes, right, if you no longer have to have your car just sitting there for you, it just opens up there's a lot more space that can be you know used in cities for other more interesting things. Uh Dimitri, thank you so much. That was a pleasure. Thank you guys. That's Dmitry Dolgov, co CEO.

Waymo. Hey, thanks so much for listening to the show this week. Please make sure to click the follow button on your podcast app so you never miss a new episode of the show, and as always, it's free. This episode was produced by Carrie Thompson with editing by John Isabella and research help from Chris Massini. Our music was composed by Ramteen Era Bluey. Our audio engineer was Neil Rausch.

Our production team at How I Built This includes Alex Chung, Carla Estevez, Casey Herman, Chris Massini, Elaine Coates, JC Howard, Malia Agadello, Neva Grant, Sam Paulson, and Catherine Cypher. I'm Guy Roz, and you've been listening. to how I built this lab.

What you see above is a preview of the first minutes. One unlock costs 3 credits and covers this episode forever: full segment and word-level timestamps on this page, plus .txt, .srt, .vtt and word-level JSON downloads, as many times as you like.