Driving Self-Driving Cars with Kyle Vogt (Cruise) | Disrupt SF (Day 2)
11:33PM Sep 6, 2018
dense urban environments
self driving cars
With that, I want to welcome our next guest. He is a serial founder and he is also CEO of Cruise. Please welcome to the stage, Kyle Vogt and your moderator Kirsten Korosec. Warm welcome.
Hi. So how many people did raise their hands? That said they wanted to.
All right, we got some work to do.
Okay, here you go. Um, all right. Well, welcome, Kyle. For those who aren't familiar, Kyle is a bit of a serial entrepreneur and CEO and co founder of cruise which GM acquired in 2016. And while you were studying, working on dreaming about autonomous vehicles, think I think since pre MIT days maybe, you did take a little break at one point, started a couple of companies, Justin TV, twitch.
Then switch back to autonomous vehicles and I think you were already, had started Cruise by the time Amazon had acquired twitch, right?
That's right. Yeah.
So are we all caught up here, is that.
I think so, I mean, I called it a break to go off and start a company and work on it for eight years, which was twitch twitch, but yeah that's what I call it. Okay, a break, yeah, yeah, a middle of this autonomous vehicle journey.
Well thanks for coming back. You were here two years ago. A lot has changed for, a lot has happened actually. For Cruise, since 2016, out of curiosity, how'd you get here today?
I did not take a cruise today.
Okay so did you take a car, a scooter?
Entourage of people with me, so I had an SUV bring me over here.
Okay. Okay. Um, but next year when you come back. Will you be in.
Most of the time I'm using a cruise to get around.
Okay. Yeah, interesting. So I do want to talk about the SoftBank investment that that just happened. But we're going to get to that in a minute. I do want you to go back in your mind to simpler time, maybe pre $3 billion flush with capital time. Um, when Cruise actually head started as a different business. And you had a totally different business model, you're going to sell an aftermarket product first. And I'm just wondering, at what point in time did you pivot?
Yeah, I mean, so let's go back to there. So you're absolutely right, when Cruise started the you know, the master plan was that doing self driving cars is too expensive and too complicated for a startup to do. So we're going to do the simplest thing we possibly can and retrofit existing vehicles with some functionality that gives them additional you know, a features or functionality that people would want and be willing to pay for.
And so we started building basically a highway autopilot system before autopilot as it exists today was around. And that was our business model. We were going to sell these kits or retrofit, you know, existing vehicles and make them really cool. Basically make them in to self driving cars with limited capabilities and use the profits from that to actually do you know driverless cars at scale, which we're doing today.
But you know, as we started going down that road, we realized this problem actually succeeding at this business is going to be a lot harder than we thought, because we did a survey of potential customers. And we found that, you know, our hope was that everyone in a certain city, a lot of people are going to have a Honda Civic or or, you know, Chevy Cruze or certain type of vehicle or something. And we'll only have to retrofit a few kinds of vehicles.
But it turned out there's like, hundreds of different types of vehicles and different makes and models and actually making the system work on each one of them was going to be really, really hard. And at the same time that that was happening, we sort of watched the rideshare companies become huge. And we knew that this sort of critical ingredient ingredient to make ride sharing really successful is is sort of the software. And today that software is human drivers.
But I think you can make it better and and less expensive and more capable version of that software with driverless cars. And so we're like, you know, we're just going to go for it. This is what we signed with to do anyway. And now that there's interest from a market perspective, maybe we can skip step one of this master plan, which is make all the profits from that early business and just go, go go straight forward. And we did.
So were you just sitting around a table and just let's rip the band aid off and make the move. And you had already taken deposits, I think for the aftermarket product, right?
Yeah, yeah, we had to give them all back. But we, yeah, it was basically that it was sort of hanging out with some friends, other startup founders. And and just, startup our founders are pretty boring to hang out with, because all they do is talk about startups. But we were doing that and we were sort of banging around these ideas and the concept of just sort of skipping that plan and going for it, you know, we were talking about that in that last to my brain.
And then within the course of two weeks, that kernel of an idea had turned into a very big one. I started doing all this research and market analysis and trying to figure out how much money would take. And basically, within the span of two weeks, completely altered the trajectory of the business, got all the investors on board all the employees and went for it. But sometimes, you know, when when a colonel like that is planted in your head, and it just starts clicking, and every single box are interested in, you can just keep checking. It's like, Oh, this makes so much sense when you have that moment. And I feel like you have to go all in.
Was their revenue, any pushback from investors when you did that?
Yeah, I mean, we, we, I made some mistakes in the early days of cruise one was doing you know what today they call like a party rounds where there's 50 or 60 investors on the cap table. And the reason for that is complicated. But basically, I was skeptical of being able to raise money for this business. And so I pitched 120 plus investors because I thought it was going to need to just take the shotgun approach and get lots and lots of shots out there.
Hope one one works and half of them invested. So I got lucky. But so I had all these investors and the odds of getting 60 people to agree on anything or close to zero, but eventually got enough of them to sort of agree to this pivot. I didn't really need their agreement, but I got enough of them interested, they wouldn't try to, like forced me out or anything and the rest of them are pretty much ignored.
Cool. Um, so SoftBank recently invested 2.25 billion and then GM through and another billion for a clean around well very unreal number actually. That's a lot of money for a private company that it's pre revenue.
Yeah, it's it's a lot of money in and as a founder I'd say who is our cut my teeth in the in the early 2000s and went through 2008 when basically startups are failing left and right and you're just like fighting for for every day of you know, capital fund Austin startup I'm completely aware of.
of the fact that this is a lot of money. But on the other hand when you look at transportation which is one of the largest markets in the world, you know, even if you do some envelope math, there's 3 trillion pastor miles driven in the US are miles driven. And
if you you know, rideshare trips are a couple bucks a mile today. So if you multiply those, the the market size is enormous. And so the idea that you could you could actually invest you know, a billion or $3 billion into a company to have a really good opportunity to win or capture or even take a significant chunk of such a huge market from that perspective it's it's almost obvious.
So you're still targeting to launch a commercial product basically an autonomous vehicle ride hailing product.
Without drivers that's the big.
Without drivers so is $3.25 billion enough to get you there.
We'll see. This is the thing about this this businesses you know, we're dealing with infrastructure
Big, you know, these are these are robots, but they're their cars. And each one might have 30,000 parts in it coming from suppliers over the world. These are electric cars, they have to be charged and cleaned and maintained. And so just the capital requirements for this business to do it right to do it safely and do it at scale are pretty high.
So, so, it's possible that you could actually go back to SoftBank GM or another investor.
I mean, our cash position today is pretty strong, feeling very comfortable with where we're at. And we're, you know, my mindset is basically, if there are ways to turn capital into, you know, basically accelerating the path of the company, making it scale faster and all those things those are things that we should be investing with capital. And so when we find those opportunities, whether that's an acquisition or you know, way to hire more people or invest in a new technology will do it.
So does soft things money give you more independence in a way from GM?
Yeah, so good question. Because, you know, a lot of people look at that deal and think, you know, it's a spin out or something. But I don't really view it that way. I think there were two key points with this deal, besides the money that I think are really important. One is we created a new class of equity, which was just written about. And that's really great. Because, you know, the trajectory of cruise the value of cruise over time looks pretty good if you were to, like, plotted on a chart and you know, all the people who are working at cruise now actually get shares of cruise so as that that business does well, even though it's a part of the GM, they could participate in the upside that's there. And so that really helps us get talent and more people.
But on the independence question, so we have that. But the goal is not to be independent, like independence is seen as in the startup world is a great thing. You want to be independent from the parent company and do whatever you want.
But like the magic of cruises, actually, that we aren't independent that we have this great partnership so that you know when when we're working on a car or we find a flaw in the way it does something out on the road or we find a durability issue it were a phone call away from the guy who designed the part and we can get it fixed and change same day without going through MBAs and lawyers or trying to figure out who owns the IP and all that kind of stuff that you have. If you have like, sort of an arm's length agreement with the car maker.
So we've done really well with that our goal is not to be independent Our goal is to basically incentivize our employees to do really well and get this product to market as quickly as we can.
So and prior previously employees were offered shares in GM and so now under the GM cruise holdings it's switched over to being able to have equity in cruise which you view as a good recruitment and retention.
Yeah, I think so. I think so. It's it's you know, that's that's sort of the magic of startups is when you're able to give people equity in something that has the potential to become very large and grow very quickly that's that's a you know exciting reason among many others to to join a startup.
So you working pretty closely with GM, even how heavily Do you rely on them for their expertise and like, what are you talking to them about realize just manufacturing? Or are there other areas where they just have the know how?
I mean, so I guess there's a number of answers that question but like a company like General Motors, very large company has vast resources and capabilities. And so it's almost like we've got this, you know, room service menu of things we can go find, it's like we need help, you know, in real estate department negotiating lease, so we can go find a person to do that in GM, or like, we need to figure out how to make this part that's vibrating and squeaking in the car. We need to like have like five engineers jump on that who have done this stuff for 100 years, know how to do it or
you know, like another thing that the GM is really good at is that validating vehicles and making sure that they come out of the assembly line basically the low defect rate and they're gonna last for hundreds of thousands of miles but also on system safety and safety is a big topic in this industry. And it should be and it's nice to have, you know, the perspective of people who have basically seen it all
And know how to approach these problems. And so there's vast resources inside of GM that we can kind of tap into. We need to, and that's, um, you know, despite the the prevalence of Silicon Valley hubris, and like this attitude that we can do everything, I think the the wiser thing that I've realized that the thing that I've learned is, you know, you don't want to reinvent everything, you want to focus on inventing the things that are core to your business. And you can do really well and uniquely well and partner and rely on others who have experts in other domains. And that's what we're doing.
So when you want to bounce an idea off of someone or you do you have like Mary Barra on speed dial or or Dan?
We talk all the time we talked 20 minutes ago but no, it's it's great to have I say leaders I respect and I think a really brilliant at the top. I mean, as you point out, I'm a serial entrepreneur and so I, I'm here because I want to be and I think this is the best place for me to be in the best between the partnership with Cruise and GM. I think we the highest probability of deploying these cars at scale and achieving all the benefits all come from that.
So you're testing in San Francisco right now see the cruise vehicles everywhere in San Francisco going to be the first market where you launch your commercial Commercial Service.
We'll see. There's a lot of factors that go into that.
Okay. And if not San Francisco, where would you launch?
Well, the the. So I'll give you sort of a taste of our master plan here, version two, which is we're going to start by deploying these vehicles in dense urban environments to begin with. And so we're testing in a dense urban environment like San Francisco. And the reason for that is really twofold.
One, it accelerates the development of the technology itself. If you want to get better, faster, you bring you bring in like the toughest situation to challenge yourself.
And the second is these vehicles in the early days are low volume, there's not going to be a million of them on day one, which means they're a little more expensive. And so to have a real business, you've got to deploy them in an environment where there are lots of customers and
And lots of retro users. And also where we can keep these vehicles busy pretty much all day long. Because, you know, basically, the second you produce them, you're trying to return the capital on that investment. And so we want to keep them carrot, you know, basically just like an airplane, it's only making money when it's in the air. So the self driving cars have to be carrying passengers all the time it's going to do that it didn't serve an environment could be San Francisco, it could be another place,
then as we ramp up the volume and add more capabilities will go to less dense environments and suburban areas, and then eventually rural,
and then scale up the volume which then drops the cost and means we can operate in more environments and so on until, you know, eventually we get every person on the planet in a self driving car and get them a safer, more affordable and, you know, more environment, environmentally friendly form of transportation.
You mentioned really, you actually think that there is a market opportunity to go into rural environments?
Yeah, I mean, so, we we have this sort of simple way of looking at things which is, you know, depending on on where you live, and how much you pay for transportation today and how you use it, there's sort of a breakeven point in cost and convenience. And when you cross over that threshold,
it sort of makes more sense to give up owning a car and use a shared. avi and, and even if you're in a rural environment at a certain point, you know,
it becomes it becomes more cost effective and more convenient to give up your vehicle. And the magic of ABS in the rideshare industry is that over time, we'll be able to beat that costs down and get it to the point where it is a better choice for you. Because unlike your car, which is sitting idle, 90% of the time, our abs are going to be busy all the time is that that that increase utilization means we can drop the cost and you know, unlike the rideshare one point O business today, where you've got humans doing the driving you're never going to get them to accept you know, a fifth of minimum wage in order to provide this service to rural areas but with a views we will.
Interesting the area that you just described it sounds like San Francisco would be the perfect environment.
There are lots of dense urban environments.
So let's pretend hypothetically, it's going to be San Francisco. Okay. I, why would you launch in such a complicated environment, you're also testing in Scottsdale, you know, right in the neighborhood, next to where way mo is why? Which is a simple and a lot of ways, a much simpler environment. So why wouldn't you just do something that you can control a little bit better?
Well, I mean, so the thing that drives me and drives the people who join cruises the potential impact we can have on the world with technology. It's, you know, like I said, safer and you know, it's sort of a secret backdoor to make the air in the cities cleaner by replacing in combustion engine vehicles with with electric ones that are abuse and so the sooner we can do that, and then the more scale we can have, the better we'll feel as a sort of a company and then also the economics are better.
And so the question that that you would ask I think, is, if I have this vision of where I want to be, which is scale. And lots of people using this great thing is is the shortest path between where I am today in this future? Is it to go test or deploy in an easy city and then try to go from there to the hard city? Or is it to go straight for the hard city?
And so we've thought a lot about this and decided that well, we could sort of plant the flag and launch in a simpler environment that's we could pat ourselves on the back for that but that's not going to get us to our end goal faster. And you know, one of the reasons we've raised all this capital is so we can go Beeline straight towards that and, you know, achieve the result we want sooner.
How many, about how many employees do you have right now?
I believe around 900, which is terrifying to me because we've grown the company 30 to 40% per quarter for the last 19 quarters and I now have gray hair.
Can't see it from here.
You know, the rides.
Um, well, I wanted to talk a little bit about talent because there's a sort of a crazy environment that's going on right now, in the world of ABS, it's a pretty limited pool of people who have really deep experience in robotics and AI. And that sort of created this environment of poaching. And these just crazy we've talked about this crazy benefits packages that have been offered to people with with that type of experience. But you've told me that cruise isn't really doing that are not participating in that sort of chasing game. So how are you able to do that?
Well, we certainly compete for talent. But I think you're right in that there's a there's a common theme right now in the industry where there are maybe 100 to 200 engineers who worked on the DARPA Grand Challenges back in 2004, 2005, where was the first I would say, serious effort that was made by lots of universities to build self driving cars. And the idea is, if you want to build a self driving car today, you have to snatch up as many of these people as you can from that early cohort as possible.
They're going to solve all your problems. What our philosophy is a little contrarian is that we're going to hire the smartest problem solvers, the best computer scientists, the best programmers, teach them the little bit of additional domain knowledge. They need to be successful as an AV engineer,
and scale the team that way. And so we have access to a much larger pool of talent than the people that are trying to dog pile on this, this early group of 100 to 200, I would say similarly, the other thing we've realized is that bit contrarian is despite the fact that, you know, one of the most challenging technical problems we do is writing the software the brain for these cars that actually, you know, navigate these complex urban environments. The way you get to that end result faster is actually by hiring many, many more engineers to build simulation capabilities in tooling and testing frameworks
and data infrastructure and all the other systems around those people working on the AV technology to make them go faster.
Interesting. I want to fast forward now 16 years in the future. You're a new Dad. Yes. And and so when your son turns 16, will he have a driver's license?
I don't think so people today like you know a lot of millennials don't even want to drive a car. So if that trend continues combined with the prevalence of self driving cars, certainly in dense urban environments and then hopefully many other places should mean that that if you want to drive a car is probably because you you want to go to a track on the weekend and and bring some rubber but that's that's a different a different reason to drive than I think most people do today.
Three really quick lightning round questions. And then we've got to wrap up on you're allowed one pass. Okay.
Great questions first?
No. Um, we're going to start with a simple coding question. Tabs or spaces.
Okay, transportation utopia, a world with only autonomous vehicles or a mix of a BS in human drivers.
Definitely only autonomous vehicles.
And are you gonna launch in San Francisco or San Francisco?
I guess I get the pass rate. Alright.
Alright. Thanks for coming.
Thank you for having me.