How I Built This with Guy Raz - HIBT Lab! Cruise: Kyle Vogt

Episode Date: October 27, 2022

Most of us are familiar with rideshare apps at this point. We tap a few buttons on a phone and...voila! A vehicle arrives to take you virtually anywhere you want to go. But what if these vehicles could operate entirely without a human driver? Will we one day live in a world where most cars drive themselves?Kyle Vogt believes that autonomous vehicles will fundamentally change how we get from place to place, and soon! After being part of the team that launched the video game streaming platform Twitch, Kyle charted a new course in 2013 by founding Cruise, which was acquired by General Motors just three years later.This week on How I Built This Lab, Kyle talks with Guy about the process of building a fleet of fully driverless ‘robo taxis’—which are now available for service in San Francisco and coming to more cities across the U.S.. Plus, the two discuss the potential of autonomous vehicles to reduce the alarming number of vehicle-related fatalities and injuries experienced every year.See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

This podcast is brought to you by Squarespace. I talk to entrepreneurs all the time who are looking for a way to upgrade their digital footprint. Well, whether you're just starting out or you're scaling your business, Squarespace is the easiest way to build a great website that stands out. It's an all-in-one website platform that gives you everything you need to claim your domain, showcase your products, and get paid. Anyone can use Squarespace's cutting-edge design tools to build an online presence that truly reflects what makes your business special.

There are templates, intuitive drag-and-drop editing, and even an AI-enhanced website builder. Then, Squarespace's built-in analytics tools help you make smarter business decisions. Review website traffic, learn where to focus engagement, and track revenue all in one place. Looking to grow your business, Squarespace even offers fast, easy business financing

through Squarespace capital. Go to Squarespace.com slash built for a free trial. And when you're ready to launch, use offer code built to save 10% off your first purchase of a website or domain. Loans issued by Celtic Bank and serviced by Stripe, all loans subject to credit approval. This show is in partnership with Airbnb. This past summer, I took my family to Vienna, and it was incredible. We spent our days wandering the old streets, stopping for coffee and pastries, visiting museums, and just soaking up the history of one of the most beautiful cities in the world. And one of the things that made the trip so special was the home we booked on Airbnb. It had tall windows, beautiful old details, and plenty of space for all of us. And being in that home on Airbnb, right in the middle of Vienna, walking distance from so much of the city,

made it feel less like a visit and more like we were actually living there. Plus, taking a trip is the perfect time to host your space on Airbnb. Your place with all of its personal touches and its amazing location could make someone else's vacation even better. Your home might be worth more than you think. Find out how much at Airbnb.ca.com slash host. Hello and welcome to how I built this lab. I'm Guy Raz. So a few weeks ago, I had a chance to go for a ride in a driverless car. And I don't mean the hands-free driving on a highway like what's available in a test. or in a bunch of other new cars right now. I mean, I got into the backseat of a car and there was no one in the front seat.

The car drove completely by itself. And this wasn't on some closed course. This was right in the middle of the city of San Francisco with other cars on the road and pedestrians, cyclists and all the distractions that any human driver might face. And it was actually amazing. The car was an electric Chevy Bolt and the company behind it is called Cruise. And right now, they have a whole fleet of fully autonomous Chevy Bolts giving rides to people around San Francisco every night. Cruz was founded in 2013 by Kyle Vote, and before working on Cruise, he dropped out of MIT to help start Justin TV, the company that eventually became the streaming platform Twitch.

Cruise was acquired by General Motors in 2016, and since then, the company has gone full steam ahead with B. building and deploying cars that can drive themselves, with Kyle in the driver's seat having stepped back into the CEO role earlier this year. Kyle has always had a love for robots and self-driving machines. In fact, in college back in 2004, Kyle was involved in a competition that was a pioneering event in the history of autonomous vehicles. While at MIT, you got involved in something called the DARPA Grand Challenge, and this actually I think is widely considered to be like a kind of a pivotal moment or maybe really important moment in the history of autonomous vehicles. What was that grand challenge in 2004? Yeah. I mean, it was put on by DARPA and their goal was, I believe, to accelerate, you know,

essentially autonomous vehicles like create the industry, get it going. I guess there's a bunch of different separate research projects going on and different universities, but it wasn't moving as fast as maybe it could. And so the government put up a prize pool, $1 million, which anyone could go after any of these colleges or private companies. And it was a challenge to see if you could build a vehicle that could drive about 70 miles through the desert following these kind of windy, rocky roads and get to a GPS destination. And at the time, this was state of the art. You know, people had done this with small robots on pretty limited terrain or smooth terrain, but having a large vehicle navigating on roads at speed was kind of at the edge of what was possible then in 2004.

So you guys got, from what I understand, the Ford F-150 was donated to you, and your task was to turn that into an autonomous vehicle. That's right. So that meant the main pieces are a computer. So we put a rack of servers in the cab of this pickup truck. then other element is the sensors. So we had LIDARs and cameras on this vehicle and wired those into the computer. And the last piece is the actuation. So some way for the computer to actually make the vehicle move. And so this is a motor we attached to where the steering wheel used to be and an electronic actuator that could move the gas and brake pedals. Wow. And did it work? No. No. So we of course being this was an undergrad led team. And so. So we had no real funding or support from the university at the time.

And so we had exactly one electric motor to turn the steering wheel. And of course the morning of the qualifying rounds for the competition, that one steering motor that we had burnt out and we were unable to compete. Wow. Okay. So that was your kind of first foray into autonomous vehicles. But you move on. You continue your studies at MIT. and I know that you interned at like iRobot, that makes a Roomba.

And your goal while you were a student was to eventually do something in robotics. Probably you're thinking by the time you're in your third year at MIT, you're going to join one of these companies and become an engineer. Yeah, I mean, I wasn't sure, but it seemed like the default path, either that or go get a job at Apple or Google. But, you know, I always had this idea of doing robotics in the back of my head, just hadn't found the path yet. During your junior year, you heard about these guys at Yale, Mitch here, Justin Kahn, and Michael Seibel. They had sold something, like a calendar software through eBay. You hear about these guys, and you think, that sounds interesting,

and what you just email them and say, hey, can I meet you? What's the story? Well, kind of. There was an infamous email mailing list in MIT called the Jobs List that was run by a faculty member in the computer science. department and it ended up being a place where Harvard and Yale business school students would basically send out a request for some engineer at MIT to kind of build their app or product for them. And so most of those were sort of, you know, half-baked attempt to, you know, frankly,

to get some kid at MIT to do the work for them. But when I saw the note come across from Emmett and Justin and then Google them, I was impressed because they seemed to have, you know, some actual experience building and selling companies. And so I was intrigued enough, especially by what they were wanting to do, which was to build this reality TV show that would stream 24-7 to the internet that I reached out just to figure out what they were up to. And the idea was we're going to have a reality show, live streaming reality show about this guy, Justin, this guy right here.

Is that what they told you? Yeah. I mean, they said, look, we have all sorts of really interesting, wacky conversations, and we kind of wish that people could experience this or participate somehow. And so why don't we just, you know, take that to the extreme and put his whole life on the internet and see what happens. And that was that was pretty much the extent of it from a business standpoint, you know, see what happens. This became Justin TV. Again, the story, we tell the story in an episode of the show with Emmett Shear. You can find it.

It was just a couple episodes back. so well worth listening to. What were you intrigued about with this idea? I mean, you were a software engineer, robotics guy. I mean, you don't strike me as somebody who'd be interested in a reality TV show about just some random person. Well, what was interesting to me was there's an interesting piece of technology that had to be built at the time. This is before iPhone. So there was no, you know, streaming from your phone to the internet like that's pretty easy to do today.

Now, of course, it's easy. You just stick on Instagram live or whatever you want to do. and you can broadcast to the world. But in 2007, you actually had to build the technology to be able to film yourself and stream that live 24 hours a day. The technology just didn't exist. Yeah, that's right. I mean, it ended up being a backpack that had all this equipment,

you know, a video encoder that would encode a video stream and digitize it. And then some software that ran on a small Linux computer that would take that video stream and dynamically allocate, you know, the bandwidth across. three different cellular modems. So basically like see three different cell phones that were in this backpack.

And the idea was, you know, maybe if you're having poor service with Verizon, maybe AT&T will have better signal there. And so you use as many cellular networks as many connections as you can and hope that you have enough bandwidth to stream live video, even if you're in the basement of a building

or walking outside. No matter where he went, we wanted that stream to stay on and remain high quality. And that was a key, part of building this always on life casting experience that Justin wanted to create. Your intention was to just kind of join these guys for a couple weeks, but you, you sort of joined them during a break, a school break at MIT, but you would never go back.

That was it. You kind of, once you were in this project, Justin TV, that was it. The whole team eventually moved to San Francisco, and you, of course, moved out with them. Yeah, Michael Seibel slyly bought me a one-way ticket. And so don't worry about it. You know, after a month, we'll get you a ticket to go back. Of course, I never ended up doing that. And I had a blast out there trying to build something with this team.

And, you know, the hardest part about that was, you know, talking to my parents about my decision to defer my return to MIT. I grew up in Kansas in the Midwest and, you know, getting a college education was a high priority thing. And so as the months turned into many months, some anxiety started to build there for sure. Yeah. I mean, I imagine, I mean, on the one hand, yes, you were building really cool technology. And that was what got you excited. But the consumer-facing part of this, Justin TV, probably got your parents thinking, what was he working on? Like, he's not going back to school to do a reality show about this guy who's filming himself all the time.

Yeah. And, you know, my dad at the time worked in a bank. And so in the Midwest. And so he's, you know, very basic questions for a business. your revenue, who your customers, what your margins look like. And the answer was like, well, nothing yet, but maybe soon. Right. Fair questions, right? Yeah, absolutely. But it did really, I mean, you guys got into Y Combinator and that famous accelerator program, which would be part of again with Cruz several years later. But really what, what of course happened with Justin TV was

it eventually became two other products. One was Twitch and one was Social Cam. And you you end up sort of focusing on Twitch. And Twitch, as most people know today, is a platform for video gamers to sort of live stream their gaming and so on. Was that appealing to you? Were you, or are you like a video gamer? And is that something that got you really excited? No, not at all, which is why Emmett was a good leader for Twitch. And, you know, I was happy working on the video streaming technology and making it so that, you know, when you have your your Xbox plugged into your into your computer, we can stream that video, it can hit the data centers, and then be rebroadcasted, you know, to millions of watchers in real time reliably and

for a really low cost. And so that was the technical problem that I was really engaged in, especially in the early years of Twitch when, you know, it was just pretty hard to do that. So, all right, you decide to leave Twitch in, I think in 2013. And why? What was there reason? I mean, it was really starting to do very well. And of course, we know what happened with Twitch. But why did you decide that it was time for you to try something different? Well, we got the technology working really well. We had this globally distributed live video network. And as I said before, I'm not really a gamer. And so I kind of lost interest. There was not much more there for me to do. And Emmett was doing a great job running the business. And so I was kind of searching for the next thing. And so what were you

thinking about. Well, I had, you know, after going through Justin TV and Twitch, which at that point for me was an eight-year-long adventure and pretty grueling. That was most of my 20s was cutting my teeth on building a company, especially going through some rough economic times. It was difficult. And I realized if I want to do this again, do another startup, and I did, I wanted to build something new, it better be something that I really care about. And to me, that meant, you know, the success is driven by the quality of the technology. I also wanted something that had a high positive impact on society. And I wanted something that I felt I could commit, you know, a decade to because in my experience and from what I've seen, it really takes a long-term commitment, something on the

order of a decade to build something, you know, meaningful. Yeah. And out of that filter that, you know, on potential ideas, I landed on self-driving cars after it, you know, I became reminded of what I had done back at MIT. And the DARPA grand challenge. And that, so you were thinking about a bunch of different possible things to do. You land on self-driving cars. This is an enormously challenging problem. But this is something that you thought, okay, it's 2013.

I'm going to put at least 10 years of my life into building this thing. And what was the sort of the product or the technology that you wanted to commercialize? Well, I mean, I guess as I started thinking about self-driving cars, the first question that popped into my head was why. Or, you know, is this worth doing? And it doesn't take long if you poke around at statistics. You know, even today, there's 40,000 people that die each year in car accidents, you know, a million people injured in the U.S. And almost all of those accidents, there's some form of human error involved. And so going back to this notion of automating the sort of repetitive or mundane tasks, like, you know, if we can chip away at that, that's a really big deal and a problem worth working on.

And just thinking about the time we all spend driving, it is a catastrophic tax on society and productivity, the amount of time we spend, you know, sitting behind the wheel. So anyways, I thought that was worth, that was definitely worth doing. There's clear benefit to society. I also knew, though, that Google had been working on this for a while, you know, after the DARPA grand challenge. And rumor had it, they had spent something like $100 million, you know, to date. And they didn't have a product yet. And so I thought if I'm going to go after this, I can't do exactly what they've done. I don't have that kind of resources.

I don't have, you know, an army of brilliant engineers from Google. So I thought, you know, maybe that what's the simplest form of this that people would be willing to pay for that I could build? And I ended up settling on essentially designing a system that could drive for you when you're on the highway. Right. And that sort of narrow version of this problem, basically keeping the car between two lane lines, seem much more tractable and also something that would provide immediate value to people. So you start working on this project in 2014. I mean, now so many new cars have a version of this kind of technology that keeps you in between lanes. If you have a Tesla, there's the autopilot system, which essentially you've got to drive the car, you've got to keep your hand on the wheel.

There are other cars that now have this technology, which is, and it makes sense. I mean, the idea that you had was you would build a kit, I guess, that you could, in theory, you'd be able to buy and then add on to your car. And then it would turn your car into like an autopilot on the highway. That's right. And we went from, you know, nothing to a working prototype that could drive from San Francisco to Powell Alto. you know, the highway stretch of that in about three months, you know, just with a scrappy team working in a garage. And that was kind of our first proof point that we could make this technology work and do it, you know, at a much lower cost than people thought was possible at the time.

And how did you raise the money to get started? I mean, just the capital cost of this thing from the beginning were going to be huge. You wouldn't be able to do this with like two or five million dollars. You had to raise tens of millions. Was that challenging or were you able to do it on the strength of which you'd already done at Twitch and Justin TV. Well, a little bit of both. So any big, bold idea that carries a lot of risk, especially technical risk, you usually have to fundraise in stages. And so you take the biggest, scariest, most formidable technical challenges and prove that you

can solve those on a small scale. And that enables, you know, that builds some confidence with potential investors that, hey, if a small team can solve, you know, a big part of this problem in a short period of time, that gives me a lot more confidence that with, you know, a larger amount of funding and potentially a larger team, they could actually go after the full problem. And that was our philosophy. If we could build this proof of concept that showed highway driving, you know, really quickly with a small team, then that would enable us to raise the funding to go for something more ambitious, like what we ultimately ended up doing, which is full driverless robotaxies.

We're going to take a quick break, but when we come back, more from Kyle Vote, the co-founder of Justin TV, Twitch, and now the driverless car company, Cruise. Back in a moment, you're listening to How I Built Welcome back to How I Built this lab. I'm Guy Raz. And I'm talking with Kyle Vote, one of the co-founders of the streaming platforms, Justin TV, and Twitch. After leaving those companies in 2013, Kyle went looking for the next big problem he wanted to solve, and he landed on driverless cars. So, all right, you start working on this in 2014, and that year you start taking pre-orders for this RP1 kit,

this highway autopilot system. It turns out that you never ended up making these kits at all, because I guess you soon after you announce that you abandoned it, and you kind of decide to focus on fully autonomous vehicles for urban environments, not these sort of highway autopilot cars. Why is that? Well, we, two things happened. One is, you know, our eyes got a little wider as we started looking at the complexities from a legal and engineering standpoint to build these kits that would work on lots and lots

of different cars. And so he thought that was doable, but it was going to be a long slog. And we'd spend a lot more time on adding support for new vehicles than we would actually making the self-driving technology better. So that was one thing. It was going to be a tough business to make it work. On the other hand, what had been happening in parallel is these ride-hilling companies like Uber and Lyft had exploded in popularity. And if you look at the economics of those businesses, the vast majority of the revenue that comes in goes right back out to the driver. And so it became obvious to us that there's a massive and growing market for fleets of vehicles

that, you know, have a lower cost than what it would cost to employ a human driver. And we thought that was enough evidence of market demand that we could do a pivot, a big bold move to basically abandon the business we had started building and recruited engineers for and raise money for and do something completely different, which was to build robotaxies. So the idea was at this point, let's build autonomous taxis or start that way in and deploy them in cities. And that will be the focus of what we do. Yeah, that's right. And, you know, the reason taxis made sense versus cars that you can go out and buy at a dealership is, you know, with the taxis, it would be a fleet of vehicles that we would own and maintain. We could

limit, you know, where they operated. And they operate, they could operate, you know, 20, 22 hours, hours a day. And so because they have that much potential to earn revenue, the vehicle itself can cost a little more than, you know, a vehicle that you would sell to someone at a dealership. And that made sense because at the time, a lot of the technology we wanted to rely on was still very expensive. The computer systems, the sensors, all that kind of stuff. And so it didn't pencil for something that you would go out and buy, but it did make sense economically as part of a Robotaxy Fleet. So as you're developing this technology, you're focusing on a number of things to make these cars operate and to make them operate safely. So there's cameras, of course, on these cars.

You use LiDAR, which is like a laser beam that is basically bouncing off the environment, feeding information and data to the car. There's radar, which does the same thing. Same thing with radio waves, and there's GPS. I mean, tell me a little bit about. the technology you you start to deploy into these cars to make them see like, you know, without a human driver. Sure. Well, I mean, the building blocks are the same, you know, even how we started a few years before. You've got a set of sensors. You have a computer that makes sense of the sensor data. And then once it's done that, it sends signals to some sort of actuators that that move the vehicle. You know, and our view early on was that this is a really hard problem. What we're trying to do with a driver

has never been done before in the field of AI. There's three big, big, big, big challenges. The first is just the complexity of driving in an urban environment. You've got, you know, people driving on the wrong side of the road, pedestrians running onto the street, trash cans falling over, like leaves in the street, debris, cyclists, like construction workers directing people with their hands. It's a very unstructured, complex scene. And if you think about what AI systems have done well on in the past, a few decades ago, computers beat humans at chess for the first time.

And a few years ago, computers beat the best video game players, you know, for some of these strategy games. But to this date, no one has really cracked a task as complicated as driving. So there's the AI challenge. Right. The second one is the fact that this is real time. You know, when you're driving at speed, this system literally has. about the blink of an eye, a couple hundred milliseconds, to process all that sensor data,

this complex AI problem, make sense of it all, and then send a signal to the vehicle. And so that real time piece means you're on the edge of what is possible with today's compute technology. And the last piece is that it's safety critical. There's really, you know,

no margin for error here. People are entrusting this kind of system with their lives. And so a lot more engineering has to go into making it, safe because unlike your phone or your laptop, it can't just crash. You know, it's actually have to handle every kind of potential failure or issue gracefully to keep the passengers safe. So you put all those three things together. This is a really challenging problem. All right. So it's 2014. That summer, Amazon acquires Twitch for almost a billion dollars. You were no longer at Twitch, but of course you had a stake in that business as somebody who had co-founded it. And you're now

working on this business, and something like a year and a half later, GM acquires Cruz. First of all, what was the – I mean, I think I know the answer to this, but from your perspective, what did GM acquiring Cruise get you? I mean, it – because oftentimes a big company buys a small company like yours and then just, you know, shudders it. I mean, Google has acquired hundreds of companies that no longer exist. And there are many, many examples. I'm sure GM has acquired companies that no longer exist. So, I mean, when GM approached you for the acquisition, obviously, I'm sure there were a lot of things about it that were attractive, the money, the resources that GM would bring.

But also, there was potential for this to be the end of Cruz. Yeah, I mean, the track record of large companies buying small startups, you know, is not very good. You know, they don't typically turn out well. But where we were is we had, at this point we had built, maybe like 30 or 40 people, we had built a prototype robo taxi that could drive around the streets of San Francisco. It obeyed traffic lights, it changed lanes, it made turns, all those kind of things. It still needed a human safety driver, but, you know, from the outside, from the appearance, it looked like it was getting close to something that could be commercialized. So we were talking to General Motors. We had been talking them for a while.

and they saw that, you know, the future of, of their business could be fundamentally disrupted by this kind of technology. And it's one of those things where you either, either like try to fight it and try to put off something that's inevitable, or you embrace it and try to figure out, you know, how you can be a part of that, that future. And in our conversations, you know, originally we, Dan Conn and I'm my co-founder, we were hesitant to do anything that would limit our optionality to work with other, automakers, like doing a deal with General Motors might mean we couldn't work with Honda or Ford or something like that. Right. But as we talked to them more, we started reflecting on our business and realized a few things. The first is that, as you said, this was going to be extremely capital intensive.

And there wasn't a huge appetite for fundraising in this space yet. There wasn't a frenzy of activity for people funding autonomous driving startups because there was no prior M&A activity, no acquisitions, and there weren't any current products on the market. And so we knew it would be hard to raise the capital we needed. And so that was a big question mark for us. Would we be able to do that or would we essentially run out of money? And the other angle was, you know, Dan Amin, who was the president of General Motors at the time, came and said, basically, look, maybe you guys can go out and raise that money. Maybe you can build robotaxies by yourself. But if you really care about making this happen and attacking those

car accidents and, you know, accelerating the safety benefits of this and other things, wouldn't you agree that with the resources and power of general motors behind you, you could do it a little faster? And I thought that was a pretty compelling argument. I mean, we slept on it a fair few nights and ultimately came back and said, absolutely, I mean, that makes sense. Let's do it. Yeah, I mean, GM is, depending on how you measure, the fifth or the sixth biggest car company in the world. Now, with Cruise as part of the GM family, what did that mean? I mean, did it sort of supercharge your ability to, you know, to develop this technology at a faster pace and at a bigger scale? Well, we got access to, first of all, the Chevrolet platform, which at the time was one of the really only the, one of the few viable EV platforms that you could use for something like this.

And that was helpful because EVs obviously have, you know, the battery capacity to run all of the computer and sensing systems we wanted. We also had access to, you know, essentially the funding we needed to build a larger team to do a lot of the hardware development and other things that were very capital intensive. So it did kind of take a lot of the restrictions off and let us accelerate the development. And we did a pretty good job all things considered making sure we could actually unlock that speed and not become kind of swallowed by the parent company. I imagine that in the first few years after the GM acquisition, you know, it was a lot of, as you mentioned, you know, deploying these vehicles on the streets of San Francisco, which is where you're based, and with a driver actually sitting in the car, and there were probably a lot of moments where the driver had to deploy the steering wheel or had to slam on the brakes. And that's probably normal. At what point were you able to test out a vehicle with nobody inside, with no safety driver. sitting in the in the driver seat. Our first time doing that was about two years ago. So I guess that'd be, I think, around the end of 2020.

So it was in the middle of the pandemic. That's right. Yeah, we did our first driverless drive around, you know, a quiet part of San Francisco at night with no one behind the wheel. And that was a rush for all of us for sure. And to get to that point, right? I mean, obviously you had to prove that it was safe. And there are regulatory hills. You needed the, presumably the state of California to give you the green light to let you do that. I'm sure at a certain point in the development of the technology, you felt or your team felt like we're ready to go. But that was probably a year, maybe two years before you were given the permission to do it. Is that right?

Well, in this case, you know, there are regulators, especially the California DMV who want to make sure that we followed best practices and taken obvious steps to make sure that. you know, we can be on communication with the vehicle and we put it through the necessary or appropriate test given how we intend to operate it. But that's not sufficient by itself. We went a lot further and, you know, safety has to be our top priority and it is. And so at that point, you know, with our testing that we had done with this system running on the road, but someone sitting behind the wheel just in case, we had collected millions of miles of data. And from that data, we have extracted hundreds of thousands of situations like close calls or human drivers doing weird things or all these strange, you know, they call them corner cases, things that don't happen

very often, but this autonomous vehicle still has to handle them properly. And turned those into simulations where we could, each time we make a change to the software that drives these cars, we could run it through hundreds of thousands of scenarios, almost like, you know, taking a driving test, but, you know, times a thousand to make sure that, you know, no matter what kind of situation the AV is in, we understand how it's going to behave and what risk it may be encountering. And then we can make a much more educated decision about whether this is ready for prime time. And that's what we had done. And the car is constantly learning, essentially learning and getting better every time.

Well, that's the thing. There's this sort of a, you know, a chicken and the egg situation here because the more more these are deployed in the field, the more examples you find of weird human behavior or other strange situations, and those can be fed back into the development loop to make the system even better. And so that's, to me, you can't do that with a human driver. You know, the average quality of drivers on the road today is just the average quality of a human driver. But with these systems, they, you know, they're already good. The ones we had today are safe. But there's no reason to me why we can't see them become 10 times better or 100 times better than human drivers.

In 2021, the state of California gave you permission to start conducting a pilot program in San Francisco without a safety driver in the car, but it was really a sort of a, it was going to begin as this very kind of small pilot program. But earlier this year in 2022, you got permission to open it up to the public in a limited form, beta form. And I should say, I've been in one, as you know, I've been in one of these cars. And it's actually amazing. It operates like an Uber or Lyft. You got an app and you can order the car and it will come and get you and you unlock the car and there's no one in there. And it will take you to where you're going. Before we get into the details of how this works, how did you land on this concept of a taxi service? Why does that make

most sense rather than, you know, trying to basically turn this into a consumer product right away. Well, it's a good question. And there's several reasons for that. The first, as I mentioned, is cost. You know, if a vehicle is operating 20 hours a day earning revenue, you know, we can afford to pay a little more up front than you would for a car that you out and buy, but it probably sits in your driveway 90% of the time. It's not providing value 24-7 like these vehicles can. The other is that we think there's a benefit to getting this technical. out there soon, as soon as it can provide a safety benefit to the community, but before it could necessarily work everywhere all the time. So, for example, we can deploy in cities

and have these drive in San Francisco at, say, 20 to 30 miles an hour max speed, even though that technology is not ready to go 80 miles an hour on the freeway or drive through a blizzard. We'll get there eventually, but we can start getting some of those benefits right now, given the state of the technology. And then lastly, I guess the other piece is keeping these vehicles in really good shape. The vehicle we're building, called The Origin, is designed to last one million miles, which is four or five times as long as the average car will. And so, you know, when this vehicle drives you and it's responsible for your safety,

you want to know that every sensor has been inspected, that the vehicle is in good working order, and that it's in great shape generally. And so we take on that responsibility and manage this fleet. So you know every time you get in one of those cars, it's safe and ready to go. And as you just mentioned, the origin is like the, it's sort of like a little bus, a little van, and there are seats on, that face each other inside. And that essentially will be designed as sort of a ride chair vehicle, almost like a bus, but a bus that is sort of comes to you. Yeah, I mean, it's in many ways kind of feels like you're using public transit, like a train or a bus, but it's door to door. And that's really important for people that are elderly or, you know, disabled or have mobility challenges.

Being able to be picked up and dropped off right in front of your house is really the only viable solution, you know, even if you would normally use public transit or other options. And so I think it's just a great solution to have that convenience and that and provide that freedom of mobility to people who really need it. We're going to take another short break, but we'll have more from Cruise CEO and co-founder, Kyle Vote, on the future of self-driving cars in just a moment. Stay with us, I'm Guy Raz, and you're listening to How I Built This Lab. Welcome back to How I Built. I'm Guy Raz. My guest is Kyle Vote, co-founder and CEO of the self-driving car company, Cruise. When I told people I was in one of these cars, the first, and you've heard this, the first thing I say is, were you scared? And what's remarkable is in the first 20 seconds, you're like, whoa, this is wild.

You're sitting in the back of this car. And within 20 seconds, it's no different than sitting in the back of an Uber. You just don't notice it. You're like on your iPhone or whatever. You're just doing your thing. You know, you're just zoned out. It's taking you to where you need to go. But as you know, and as probably lots of people listening, though, there have been mishaps, right?

And this is normal with any new technology. But there's been a big focus and emphasis on mishaps, a car that was, you know, the San Francisco police tried to pull a car over and it kept driving. You know, another one. It did pull over. It did pull over. It did pull over. Okay. A group of cars that blocked an intersection for some time.

There's a lot of attention that gets paid to those mishaps. First of all, what's your reaction to that? I mean, do you get frustrated? Are you like, God, no one's seeing the good side of this? Or do you understand why people sometimes focus on those things that? cause disruption. Well, look, I think these vehicles are very clearly, you know, autonomous vehicles. They have sensors on the roof. They say cruise on the side. And there's no one in them. So they draw a lot of attention. And anytime they do anything that perhaps, you know, if it were a human driver,

you would not even really notice, it makes the news. And that's to be expected. But what I think is really important is that we not lose sight of the fact that, you know, we have a real problem today with car accidents. You know, I have a four-year-old son that's on my mind all the time. I have a grandfather. I never met from a car accident. And I think as a society, our goal should be progress, and we've fallen behind. And I think that self-driving cars are the only thing that I've really seen as a potential solution to car accidents because humans are always going to be human. They're always going to make mistakes. That's part of being human. These AVs, Autonomous vehicles, you know, they're safe. Our track record for safety is excellent, but then, you know,

they won't make zero mistakes. And I think they'll make far fewer than we as humans do collectively. And so I think there's a distinction between things that look kind of odd that AVs might do versus the real safety impact that they can have, you know, currently and over the long run. And so even these, you know, rare things that the people have seen or have popped up in the media, they're very rare. Yeah. Cruise is driven 400,000 driverless miles and only there's been a few handfuls of these types of things pop up and, you know, nothing, nothing serious as when it comes to safety, which is, you know, the thing that's most important. But essentially what you're saying is there's, it's impossible to say that there will never be a single incident. Even once the technology is close to perfect, there will be incidents, maybe even accidents in the future.

I think the goal here is to make something that's much safer than humans are and ever will be. And that's what we need. That's what we are in dire need of. Tell me a little bit about what's happening in San Francisco right now. I guess basically between the hours of sort of 10, 1030 at night and 5 or 530 in the morning, there are roughly 100 driverless cars driving through the streets of San Francisco waiting to get an order or somebody to order one to get a ride. Yeah, that's right. So the hours are limited now. I think in a few months it'll be dramatically different.

It could even be 24-7. But yeah, it's on the order of 100 cars. And in the evening, they all start off in parking lots, powered off. A couple people run around, flip the switch, turn all of them on. And then within minutes, one after the other, they shoot out. out of this parking lot completely empty, and they start scattering around the city, getting ready to pick people up and take them on rides. And why did you guys decide to start in San Francisco?

Because you're based there or because there are other reasons that you want the cars to be in that city? Well, both. The first thing to realize is that we're operating a fleet of vehicles and to actually turn this into a business, which is our goal here ultimately. I mean, we can't provide this benefit to communities for free. We have to have to be able to earn some revenue from it to pay for our business. operations and you want to fish where the fish are. And so San Francisco is among, you know,

a handful of major markets in the U.S. where there's existing demand and willingness to pay for ride hail services. So it's one of the obvious first places to start. It's also a really difficult driving environment. It is a major city. It has high pedestrian and cyclist density. It also has hills and fog and many other things that make it challenging for a self-driving system to do really well. And so our attitude was, you know, if we want this system to improve at the fastest rate and be ready to go in other cities as quickly as possible, we should subject it to the most challenging environment so that, you know, we know if it works here, the path to to unlock other cities and deployed in other cities is going to be pretty straightforward.

Your vehicles use radar and LIDAR and GPS and cameras. And Tesla, for example, which of course is working on its own autonomous vehicles and has a version of it, the full self-driving car, which is in beta, and many Tesla users have access to it, they only use cameras. And their argument is that LIDAR is unnecessary, that it's expensive and that actually it can't tell the difference between a floating plastic bag and a speed bump or, you know, or a dog running across the road. I mean, what do you make of that argument? What's a good question, and I think a lot of attention gets put on this debate around LIDAR, but if you don't mind, I want to take a step back. I think it's kind of a wrong way to look at it.

So our philosophy is in our business is we operate robotaxies. And so it only works if they don't have a driver. And so we have optimized for vehicles that are well equipped. They have all the sensors they need to do that job, like literally no driver, which is a very hard task. And that's where we start for our business model. Our goal, which I believe is also, you know, the goal of Tesla and other companies, is to eventually get this to be low cost, you know, works everywhere and doesn't need a driver. And if you look at the business model of Tesla, they're just starting from the other side, which is these vehicles aren't well equipped and they don't yet work everywhere and they still need a driver, but they're starting on a lower cost point.

And they're hoping that, you know, maybe they'll build software in the future that will actually let you take the driver out of the car. similar for us, we already have the drivers out of the car, but we're hoping in the future we can make it cost much, much less. And so it could eventually be on a car that you could go out and buy. And so for LiDAR and cameras, it really just comes down to cost. Does your business model enable you to put certain types of sensors on the vehicle? For us, we can do that. And I think in the future, you know, the cost of LiDAR is coming down to the point where they basically cost the same as a camera. And so I wouldn't be surprised if in the future you see Tesla's with very low-cost lighters, and you might cruise vehicles with very few, if any, lighters. And I think that's just the natural

evolution as we all work towards these low-cost vehicles that work anywhere and don't need a driver. One of the things that Tesla says about its vehicles is that it has so much data already because there's so many Teslas on the roads all over the United States that they have this neural network that enables their vehicles to be, to have better information. Is that, in your view, a fair point, given how many of those vehicles are deployed on the roads? If you're using machine learning in these systems, which most of us are, there's definitely a benefit to having diverse and large data sets.

But, you know, the data itself isn't enough to make these systems better. It's actually the ability to analyze that data, extract insights from, it and then use it to improve the product. And so you've got Tesla that maybe has a large amount of low quality data. It's just camera data from a bunch of vehicles. And then you have crews and others that have a lot of very high-resolution, high-quality data from these LIDARs and radars and other sensors. And we found that we're not really limited by the data we have, even though we've

driven maybe a tiny fraction of what Tesla has. It's really the ability to turn that data into insights and improvements in the product. Kyle, there are lots of companies working on this technology. I mean, is this sort of a race to victory, or are all of these companies going to, you know, going to carve out their, you know, they're part of the market in some form? I mean, are you going to be, do you think in the future you're going to be seeing Waymo technology and cruise technology and all these other companies in their technologies deployed in a similar way? Or do you think there's going to be one or two winners? It's hard for me to say. What I will say is we've seen already that the quality of the driving experience and the end-to-end experience, you know, from when you pull out your phone to when the car shows up, has a very large influence on, you know, whether you use that service again.

And, you know, doing that well, every touch point with the customer, whether it's, you know, the way, the text that you put in the apps, the way that you communicate with them when they want to change the destination or whatever it is, there's a lot of nuance to getting that experience right. outside of the core driving quality of the vehicle. It's not the case that all self-driving cars will feel equal when you're inside of them. I think people are going to have a natural preference. Our view is that the handful of companies that do this first and start these cycles of iteration and learning are probably going to end up with the best products. And it's likely that it becomes harder and harder to catch up as these other companies reach scale. So I don't think this will become a commodity, but I also can't say that it will be one company. I think likely a few will make it, you know, and end up building a product that is competitive with the others on the market at the time, priced appropriately, and delivers a really great experience.

But I guess you would argue that one of your competitive advantages is that you can mass produce these cars because GM can mass produce these bolts with the autonomous technology built onto it. Right? And they can do that on the factory floor. Yeah. So, I mean, that's where the General Motors thing comes in is general motors is really good at the things that Cruz as a company is not, which is, you know, the hundred years of institutional knowledge on manufacturing vehicles at scale. But it's more than just building a lot of vehicles. It's also building vehicles that last a really long time. So tell me where this is headed now. Now you're in San Francisco and you're about to expand to Austin and Phoenix. And anyone can download the app, right? And then you can get on the way to. list to get a chance to use it. On average, how long does it take to get off the waiting list? We have a lot of demand right now. Demand far outstripped supply, so it can be a while. But, you know, for the longest time, for seven or eight years, our focus was just getting the first vehicles

to operate as robo-taxies without drivers and doing all the work to make these products safe and, you know, handle any kind of thing. It could go wrong and literally anything. There's a lot of work that goes into that, whether it's a sensor malfunctioning or computer crashing or even someone damaging the car from the outside, no matter what it is, the vehicle has to behave responsibly in those situations. Now we're at the point where we have the robotaxies out there in a city like San Francisco. And so the technical problems, the software and AI problems are no longer the biggest bottleneck. Now we're turning our attention back to manufacturing and scaling and building out charging and cleaning and maintenance infrastructure in cities.

And so you're going to see us pop up in multiple cities and with more cars and more service area pretty rapidly over the next year or two. So now it's going to Austin and Phoenix. And I mean, you can operate like an Uber or Lyft now without a driver in, you know, almost anywhere. What is the holdup? Is it getting government regulators to, you know, to kind of open the floodgates and allow this to happen? Is that the main holdup at this point? Well, to be frank, there aren't a lot of holdups now, which is good news. In California, we've had to get six permits from a number of different regulatory agencies

to do what we're doing now, and we'll need even more to expand. You know, in other jurisdictions and other states, there aren't as many requirements. In some states, there's none at all. And so, like I said, we're turning our attention now to spinning up manufacturing, which is something that, you know, it's not something that, you know, it's not something that carries great risk in terms of the timing for that. It just takes a little time to spin up. So I'm pretty optimistic about the plans for expansion, and we don't see huge bottlenecks in the process. Now that we've built the core technology, started to build up a really good track record

around safety and operated in a major U.S. city. There are fewer and fewer barriers that stand in front of us at this point. So if you're looking now ahead, right, let's say 50 years from now, do you think it's fair to say that every GM car produced in 50 years from now will have autonomous technology on it that will essentially enable every single one of those cars to drive on their own? Absolutely. And beyond that, I'll say in 50 years time, a couple things will have become apparent and likely have occurred. The first is that I have very little doubt that self-driving cars will be many, many, many, many times safer than human drivers to the point where it starts to look a little reckless to let humans drive cars in places where there might

be pedestrians and cyclists. I also think for that same reason, we'll look back at today this time period as being essentially a dark age where we as a society somehow turned a blind eye to the carnage caused by people driving cars in the U.S. We take it for granted today and we kind of ignore it because there's no viable alternative. You can't, you know, just not take your kid to school or not go to the grocery store or not go to work. You have to do these things. And as a society, we're taking on a great deal of risk by putting humans behind the wheels. And so I think 50 years from now, we're going to look back and say, what were we thinking? Kyle, thanks so much. Thank you.

Hey, thanks so much for listening to How I Built This Lab. You can follow how I built this on Apple Podcasts, Amazon Music or wherever you're listening right now. The next episode is available right now, or you can binge all of our episodes, ad free by subscribing to Wondery Plus in Apple Podcasts or the Wondery app. If you want to follow us on socials, we're at How I Built This on Twitter and Twitter, and at guy.orgas on Instagram. This episode was produced by Catherine Seifer with editing by John Isabella. Our music was composed by Rumtin Arabley. Our audio engineer was Neil Rauch.

Our production team at How I Built This includes Alex Chung, Casey Herman, Carla Estevez, Chris Messini, Elaine Coates, Josh Lash, J.C. Howard, Liz Metzger, Sam Paulson, and Carrie Thompson. Our intern is Susanna Brown. Neva Grant is our supervising editor. Beth Donovan is our executive producer. I'm Guy Raz, and you've been listening to How I Built This. I hate driving. I can't wait. I want to like put a bed in the back and just sleep.