The Man Making Self-Driving Trucks
57:43 Share

On September 28th, the Global Citizen Festival will gather thousands of people who took action to end extreme poverty. Join Post Malone, Doja Cat, Lisa, Jelly Roll, and Raul Alejandro as they take the stage with world leaders and activists to defeat poverty, defend the planet, and demand equity. Download the Global Citizen app today and earn your spot at the festival. Learn more at globalcitizen.org.com.

On September 28th, the Global Citizen Festival will gather thousands of people who took action to end extreme poverty. Join Post Malone, Doja Cat, Lisa, Jelly Roll, and Raul Alejandro as they take the stage with world leaders and activists to defeat poverty, defend the planet, and demand equity. Download the Global Citizen app today and earn your spot at the festival. Learn more at globalcitizen.org slash bots. It's on!

Hi, everyone from New York Magazine and the Vox Media Podcast Network. This is Convoy, good buddy. Just kidding. This is On with Kara Swisher and I'm Kara Swisher. And I'm Naeem Araza. And I do not know what Convoy, good buddy is. What is Convoy? Yes, young person. It's a movie about trucking. There's a whole bunch of them. Okay.

It's not because I'm young. It's because I don't know about trucking. I think it's probably you never saw them. It's like you never, they were the 80s. I think it was the 80s. And it was a series of movies that sort of celebrated the truckers, including Smokey and the Bandit with Burt Reynolds and Sally Field, which was one of my favorite series of movies. And it's all about how truckers, this romantic version of truckers. Everyone loves a good truck, Cybertruck, the truck love. My son certainly does. He calls them gucks. Why gucks? I don't know. He can't pronounce truck.

He's one and a half. That was Solomon. I was thinking Louie. I was like, hey. Louie can pronounce the word truck. Yeah. Well, yes, my five-year-old godson loves trucks too. He points them out everywhere. So it's car week here at ON. And you are in San Francisco, an epicenter of innovation. You know, I'm a city gal. I'm into public transport and walking.

But, Kara, you're a car gal. Yes, I am. Now I have a hybrid and an electric car. You have Chevy Bolt, electric Hummer, Kia SUV. You're like the Goldilocks of cars. Yeah, I change cars a lot. And now I'm spending a lot of time thinking about electric cars and autonomous cars and autonomous vehicles. And also moving about a city. How do you change a city from sort of a car-centric city to one that's not? Today we have an interview with Chris Urmson, the CEO of the autonomous driving company Aurora Innovations.

And on Thursday, we've got an interview coming with Takedra Mawakana, the co-CEO of Waymo, which was formerly known as Google's Self-Driving Car Project and also happens to be Urmson's old stomping ground. Yeah. These are Kara Swisher's selected interviews. You wanted to do kind of a beyond Tesla and broader innovation interview.

in the car market. So talk about that. Yeah, there's a lot going on. There's two parts of the future vehicles. And one is autonomy, which is it drives itself, which I think got a little ahead of itself in terms of the ability to do so, although there's been great strides. And then there's obviously electric, which is a big push by a lot of, because of climate change and everything else.

That one I'm more familiar with. Yeah. And so it's how do you electrify cars? And obviously Tesla's the absolute pioneer here and continues to be the leader. But every other car company and new car companies have gotten into the act. We're at a consumer inflection point of people wanting to... I've had so many people interested in electric cars now. And it wasn't a small market, but it wasn't big. But it's still very expensive to get this thing going and making this shift now.

from fossil fuels to electrification. Yeah. Part of the push around EVs has, as you're saying, also been driven by the government. People like Governor Newsom's bill, right, when all new electric cars by I think 2035 have to be electric. And there has been a consumer inflection point in part because of the IRA, but there's also a lot of confusion, like those $7,500 rebates that people thought would be on all EVs but were only on EVs where the batteries were made in America, which were not many.

And then how do you do this at scale? You know, everything takes investment. And again, we're at an inflection point of where we're going. I don't think people are short term in the idea of what's happening here. You know, Tesla does lead the way in cars. But what's interesting over the last month, Tesla's shares, which had been really low, dropping rather precipitously, largely because of Elon Musk's antics over Twitter and investors worry, are up to $212 a share.

making Elon Musk the world's richest man again. For a long time, Elon was seen as kind of the godfather of both autonomous and electric vehicles, both those trends you pointed out. He's long made these bold predictions about having fully autonomous cars by 2017. Obviously, that did not happen. But earlier this year, he hit a big obstacle of having to do a massive recall of over 350,000 vehicles.

of those vehicles with Tesla's full self-driving beta, the FSD, because it, quote, may cause crashes. And they were pushed by regulators who were kind of scrutinizing the safety. So explain what went on there. With electric vehicles, everyone had range anxiety. Explain the safety anxiety that might come to bear with autonomy. Yeah. The Tesla recall, they had problems with autonomy. And so this is going to happen. And they usually make software updates. This is how they do it. Mm-hmm.

And I think they did weird things like stop suddenly and like when it sees things. And so Elon's different than everyone else because he only believes in a certain way to make an autonomous car. And he believes...

certain technologies, I'm not going to go into all of them, are worthless. And so, you know, he's going to see challenges. This is a really difficult problem. It just is. It just is, no matter how you slice it. This is an area where I think Elon's been really out front on, but he does make promises that he doesn't keep, right? Yeah. Which is because he's a salesman also. I still freak out in like LA when you see those little delivery things and San Francisco when you see that. You're not freaked out by people driving? I'm freaked out by people driving. Well...

I think people are terrible drivers in general. And that's where all the accidents will be, is between humans and these autonomous cars. Well, yeah, they say that. The interesting thing is that I think that humans are dangerous on a singular basis, but just like electric, there's a grid. The whole system could be dangerous with

autonomous vehicles, which is cyber hacking. I guess, yes. Taking control. They can do that now. There's a systemic risk. There's all kinds of electronics in your car right now. This is the way it's going. It's not one where we are going to have electric...

And then autonomy and new ways of doing public transportation that goes right to where you want to go at an affordable price. That's the goal. And then to keep them out of main cities. And one company can't make everything. Obviously, there's going to be competitors. And so in all the aspects of vehicles, eventually, you know, but trucks are sort of this quiet little area that's a huge business and making them autonomous is probably.

There's probably going to be just autonomous trucks in the future. I feel like we're always getting these emails from Chris Urmson asking you to get in a truck and Chris Urmson's team. I've been seeing these emails for three years. So explain how you know Chris Urmson. He's so keen for you to get into the truck. I have been talking to him for a long time. He was the person in charge at Google in self-driving, the self-driving unit.

And I did a great video, we'll put a link in the show notes to it, to me riding in their new self-driving car. And we introduced it at my conference, Sergey Brin did, in 2014. And I talked to Chris himself a couple of years later. And so I've always been interested, he's always been out in the forefront and one of the

really, he's sort of the opposite of Filon in a lot of ways in terms of being very much like, let's wait and see. We're trying this and very measured in terms of where it's going. And I've always appreciated his intelligence around the area. He's not a hand-wavy guy. But the thing that's going to be in Aurora's way a little bit will be workers. Truck drivers are not thrilled about losing their job potentially to automated vehicles. This is obviously a powerful contingent of people. Look at

the, you know, Canada truck protests over COVID, the truck blockade in Brazil after Bolsonaro lost the election or closer to home after recording this interview with Irmson, the California State Assembly voted to ban driverless trucks from the state by 54 to 3. And that bill is now headed to the state Senate. So do you think that

will slow down the move to autonomy? Yes, of course, because they have an idea that this is not coming. I'm sure the people who had wagons did this too. They did, and manufacturing, and everybody else. It doesn't matter. There's not enough truck drivers. This is true. That's the argument of the industry, and they're correct. There's not enough truck drivers for these companies. People aren't going to. It's a very dangerous job. And again, it

Trucks will do this better. I mean, autonomous trucks will do this better eventually on some routes. And it's not completely replacing people because they'll be moving these trucks around. There'll be maintenance. There'll be all kinds of jobs about dealing with them, you know, and running them. And so I think probably the last thing we'll have is autonomous planes because that makes people nervous. But planes are autonomous, actually. Yeah, they are. But they have humans sitting there as a failsafe. I think that's the question is, does the human still sit there as a failsafe? Maybe. Yeah.

Probably not. Autonomous trucks don't get tired. There's no question that it's going to happen. It's for businesses. The California State Assembly banning driverless trucks is stupid. It just is. I wish that had happened before the interview so we could have talked to Grimson about it. 100%. All right, let's take a quick break and we'll be back with the interview with Chris Grimson.

I want to start, actually, we go back a long way. You were at Google where you were working on the self-driving car before leaving in 2016. You were the first person I talked to about self-driving cars, I believe. That's probably about right. Yeah. I've been at it for 20 years now. Yeah. So explain why you got into it. Just so people don't know, I went out to Google in a parking lot there in Mountain View, I guess it was. And you had this tiny little clown car.

Quality car, come on. Okay, it was a quality car. It was a tiny little car and I ran around with it, tried to run you over several times and it didn't do that. I'm relatively happy about that. Good. So as we said, it would have been ironic, but still talk about how you got into that because I remember you were, there was a small group of people at Google, which was one of the first companies to focus on this. Yeah, so I'm a little ashamed to admit that it's,

because I thought it was cool. So back when I was a graduate student at Carnegie Mellon, there was a competition that the Defense Department put on called the DARPA Grand Challenges, which were these races to make a robot drive across the desert between Los Angeles and Las Vegas. And the idea of having a vehicle drive 50 miles an hour across the desert and kind of stay in the trail and dodge tumbleweeds and whatever just sounded really fun.

The last of those competitions was about driving around on the roads. And at that moment, it was clear that this could have a huge impact in civilian life. That if you look at the around 40,000 Americans that die on the roads every year, we can do something about that. We can give people better access to transportation.

and do that while working on a really interesting problem, do that working with great people and do something that will have a huge impact. So you went to Google of all places. Why is that? Because Carnegie Mellon's known for some of this stuff and other places, but why did you start that at Google? And talk about starting that because Google at that time would start just about anything. They had like airships, which they're back at, I think. Back when I joined there, they didn't.

which was really interesting. So I'd just become a professor at Carnegie Mellon. That was kind of up to that point, my life ambition, you know, to become faculty somewhere. And Sebastian Thrun, who was at Google at the time, had helped found what's now Street View. We had been competing in these challenges and we thought about like, hey, we should get together and start a company to build self-driving vehicles. We didn't actually have that word for it back then.

And apparently Larry told, Larry Page told Sebastian, "Why don't you do that here?" Right, this is the Google founder, one of the founders. This was Google co-founders, yeah.

And at the time, you know, Android hadn't been released yet. Chrome wasn't a thing. It was really a search engine. And so it seemed crazy to me that this company would want to do it. But it seemed like an opportunity to come out to California. The goal initially was to drive 100,000 miles on public roads and then drive this 1,000 miles of very specific roads to see if we could push the technology forward. And it was a little bit like the Dread Pirate Roberts where they said, you know, you've got

"Two years to do this and we'll probably fire you "at the end of it, but give it a go." And then it just took off from there. And we did that in about 18 months and then we continued to push the technology forward.

Eventually you got a chance to ride in one of our little prototype vehicles and went off to the races with it. So you had that prototype vehicle, which was driverless. It was like a ride at Disneyland. I think I described it like that. And drove in circles and it would stop and it would see you and you used all kinds of lighter technology, all kinds of technology to do that, sensing technology.

So why did you leave then if they were doing it? Was it that they were not going to make cars or this was Google was it did a lot of experimental things? Yeah. You know, Google was an incredible place to work. I felt incredibly privileged to have been there. It was amazing people to work with. But at the end of the day,

I didn't believe that we were going to work in partnership with others and see this technology really happen at scale. And as the person leading it, if I didn't believe in what we were doing, then I need to get out of the way. There was too many great people putting their energy into it. The company was being so generous in enabling it. And if I wasn't bought in,

then you can have three opportunities. You can either get in line, you can try and fix it, or you can get out of the way. So you went off and how did you get into trucks? Why trucks? I mean, because you were in cars. The idea is... And that's where a lot of the attention got also on autonomy and not electric, although most people think it should have been electric before autonomy, which it's moving towards, I think. It'll ultimately be both, right? These are parallel technologies that have a huge positive impact from both of them. So for us, we wanted to build a driver that would serve all kinds of markets.

And we started with cars, honestly, because- And I saw one. I just saw you had a Toyota. Yeah, we've got a Toyota Santa here and we've used a variety of different vehicles over the years. I think we've integrated onto eight different vehicles at this point. But it turns out that cars are, A, a lot easier to operate. You don't need a special license. They use less fuel. It's just easier.

But B, really importantly, when we started the company, we didn't think anyone could see far enough to solve the problem. To be able to drive a truck down the road, you need to be looking 400 meters ahead.

Basically, the LiDARs that are out there in the market today, with the exception of what we've built internally, could only see about 200. And so we spent the first couple of years of the company trying to find a technology that would enable us to see far enough that we could go do trucking. So when you're talking about that, for people who don't understand, far enough is that on a highway, you have a longer range thing ahead of you because you're going faster and at greater speed. And so you need more technology.

Yeah, well, and also if you're driving a truck. So if you're driving a car and you think about how far it takes you to stop from 70 miles an hour versus 25 miles an hour, 25 miles an hour, you'll stop at a couple of feet, right? 10 feet or something.

You'll take you 100 feet, 150 feet to stop from high speed. And if you've got a truck... With all kinds loaded up. With 70,000 pounds of goods in it, you're going to take hundreds of meters to stop. But why trucks? What was the thing? You were in cars. I was surprised when I heard you were doing trucks. We thought trucking was the right answer because for a bunch of reasons. One is...

that as a business, it was a really important business. That we've just went through this big supply chain challenge that we saw in the US. A part of that was just, we don't have enough drivers, right? There's a shortage of 80,000 people to drive trucks in the US today and we expect that to go way larger. We thought the business would make a lot more sense that, you know, it turns out we value driving a truck about three times as much as we value driving a car.

As a business. As a business. Well, as how much you pay for someone to do that job. And so if our job is to be the driver and early on, like anything, it's gonna be more expensive than over time. Well, why don't we go work in the place where the economics make sense off the bat? So you could actually build a business and you can go out and have the impact.

The opportunity to improve safety on the road, the opportunity to help people move from jobs where they're away from home for a vast majority of the time, where they die at a rate that's 10 times higher on the job than the average American. This is a truck driver. This is a truck driver. So like there's all kinds of flow through. You're helping them. They don't think that. Well, so I think it's really important to look at truck drivers and say like, this is an important job, right? You, I, we wouldn't eat, we wouldn't have clothes without people driving trucks.

But it's a really dangerous job. And it's one where we could have them with the way we think of the oral driver coming to market, doing these long haul trips, which are really not the ones people enjoy driving the most, will allow those drivers to then drive local haul.

And so they're home at night, right? They're driving in more familiar environments, it's safer. So we see that as a real opportunity. And by the way, there just isn't enough of them. And we'll get to the economics of these companies, it'll be cheaper. So let's talk about how it works so people can understand and let's try not to be too geeky here. In early April, you announced you'd achieve, it's quote, "feature complete." Explain what that means and what you're working on now. Yeah, so we've been building the Aurora Driver, which is this thing that allows the truck to drive safely.

And when we hit feature complete, that means now we have all of the pieces in place. So they don't all work well enough yet, but they're all there. So we can do all of the different kind of driving maneuvers. We can detect when the weather's bad that we shouldn't be driving. You know, we can even understand when we got into a collision or somebody collided with us and take reaction to that. And so that's all in place.

And now what we're doing is basically going through the exercise of both refining and improving parts of that and doing all the testing to build confidence that the driver is safe enough that you could have it on the road without people. Without people. So explain what the pieces are very quickly. There's...

There's LIDAR, there's... Explain the critical pieces of this feature complete. Sure. So as you think about the vehicle, first you need to be able to see the world. And to do that, we use a combination of different sensors. We use LIDAR, which is like radar, but with light. Radar that we're kind of familiar with, with the police using spec speed and cameras. And so those allow us to see out to a few hundred meters. We can see the color of traffic lights. We can see vehicles, where the road is, etc., etc.

Then we have a big computer which takes that in and then our software runs on that, our algorithms, to figure out what's happening in the road around it, how the world's going to react when we do something, and then that's what we should do to get safely where we're going. And that computer actually is, think of it really almost like three computers. So there's two of them that are redundant so that if one of them fails, we can fall over to a second.

And then there's another one, which in case both of them fail for the same reason, we have this other computer that can take over and do very simple things to make sure that the vehicle is safe. Yeah. Stop the car. So you don't anticipate major changes in the system? No, we kind of got the pieces in place to get to operating without a person on board. And then, you know, obviously over time it'll evolve and we'll make it less expensive. We'll make it able to do more things.

but over the course of the rest of this year, we expect to get to the point where it'll be good enough that it can operate without a person on board. So one of the things we're developing right now is most of the time,

Driving on the freeway is boring. But every once in a while, somebody cuts you off in a way that's really... All the time. Kind of really dodgy. Just now on the way here. We run through, we've got tens of thousands of tests like that where we can see what would the Aurora driver do and make sure it's doing something that is safe. And that's a level of testing that no human driver has ever put through.

It's years and years of experience, something that just wouldn't be practical. And so as we build that and a variety of the other testing work that we do, we get confidence that we're going to be safe out there and that Hugh's going to be fine on the road as we're driving along. We went to see in your lab here, this is a Peterbilt truck. It can be installed on any truck, but you have to...

You were talking about it in the cab. That's right. So you can think of it as it's a modular system, but you can't just go buy a truck off the lot and bolt it on. We work with our good friends at PACCAR, good friends at Volvo Truck, good friends at Toyota. Those are the big truck makers. Yeah, and Toyota, obviously, biggest car manufacturer on the planet. We work with them.

to make changes to the vehicle so that it has all the safety systems it needs, whether it's redundant steering, so that if some part of the steering system breaks, there's another part that can support, or redundancy in the braking. So if for whatever reason our little brake actuator fails, there's another one there that can take over. And so by having that and doing the work with their engineering teams, we can have confidence that

When the truck is out on the road, it'll drive well when everything's good. And then if something breaks, that we can detect it and do the right thing in that moment, whether it's switch to the backup steering system or use one of the other LIDARs instead of, you know, a camera or something because we have that redundancy. Something breaks or doesn't work, as with a plane. It reminds one of a plane. And it's for the same reason that, you know, that you...

If you're up in the air and the thing that keeps you up in the air fails, it's a bad day. We wanna make sure that as we're driving along the freeway around other people on the road, we've been thoughtful, done all the work and analysis, put all the components in place so that if something breaks, then everything's gonna be okay. Presumably other people is the problem. Most car people tell me that humans are the big problem. Yes, right. We see like people drive cars

pretty well, but humans are fallible, right? And they learn nothing. And I think that's one of the, honestly, the opportunities that we see with this technology is one, it never gets distracted. So it isn't frustrated because it had a fight with a loved one or at work on the way home. And so it's thinking about something else.

It doesn't get tired, it's gonna do the same thing over and over. It has some superhuman powers, right? The fact that it can look all the way around it when it's driving. So I find it honestly really terrifying to make lane changes on the freeway 'cause you're looking in front, but if you wanna be safe, you have to look over your shoulder. And if that vehicle in front hits the brakes in that moment, it can be a real bad day. Whereas our system is looking beside, behind, around,

always, all the time. All the time. And so those, that means that we'll end up being even safer. So right now though, the person is in a truck now. And why is that? And it obviously won't always be the case. You wanna remove people from the truck 'cause the steering wheel remains in place. It could go to manual or whatever version of that it is. And so today we are hauling goods for customers every day, right? Where, you know, FedEx is shipping stuff through us. We work with Schneider, Werner, a bunch of these big trucking companies.

And we have operators on board because we're not done yet. And so they are there to observe the system, to provide feedback to our development team, to say, hey, it did this and it could have been a bit better if it did that. And frankly, to keep everyone safe as well. That's part of their role on the road. But it's really about that development process. And then as we

get the confidence through all the testing work we're doing, all the analysis we're doing, we'll ultimately get to the point where we operate these vehicles without them on board. Without them on board. You only, though, building the driving system and not the truck itself. Others, say Tesla, some others, Lucid, you know, they're more partnering. But you were just talking about you wanted partners. Why did you do it this way instead of building the truck and selling the truck? Yeah, we thought it made sense, right? Like, you know, I try to be practical. Like, we have...

in what we do well, but we have the humility to understand that other people do things really well too. Did you use the word humility in Silicon Valley just now? I know, there's probably gonna be somebody at the door. I know, you're arrested. No, we really thought that there's these companies, they've evolved over a hundred years, right? And that means they're good at what they do. So why don't we work with them? Why don't we do what we do? They do what they do. Find folks who have a similar experience

of how this can have a huge positive impact and do that. And it'll allow us to have a bigger impact and allow us to help our partners grow their businesses, allow their customers to grow theirs. Also the money to create your own trucks, correct? We also don't wanna do, yeah. And just from a practical business perspective, why should I invest that money? There's a place where, if I think from a business lens,

the highest return on investment is where we're really good at what we do. Which is software. Which is in the software, the driving technology for this. And so let's do that. And then presumably you can take that technology and put it in cars. Now, lots of car companies are in partnership with lots of things or they bought companies.

I'm sure you've had a million offers to run some of these big car company. Yeah, there's been opportunities and what we're seeing is like any other new industry, a lot of the efforts are actually flailing and failing because it's really hard. Like any new technology, people look and say, "Oh, that's really cool. I'll make a mint there by doing that thing." Without really understanding what it takes to do it. And the way we've built Aurora is like, hey,

This is gonna happen. There's gonna be a lot of consolidation. It's gonna be a lot of failure. Let's make sure we've positioned the company to come through that with strength. That means not being overly ambitious and just build trucks. It's like, no, the driver will work with people who make trucks. That makes us more capital efficient. It's let's make sure we- Why have they failed? I mean, again, you've been, I'm assuming you were offered running every single one of these things. I was given a few options. Such as? I'll let that be. But I've had a few options and- Why did you say no?

because I thought the right way to go about this was to have this be the main bet that that this was a hard problem it wasn't one that was going to be solved on a shoestring budget it wasn't one that was going to be solved as kind of the side bed it was one that

You had to have people around you who were in the same boat with the same mission. Because at some point, just practically a company, if this is, you know, the core business, like take Ford, for example. Ford's about making vehicles for people to buy. And they had partnered with Volkswagen and owned this company.

startup in space called Argo AI. The challenge for Argo was at some point, either the management changes and was like, you know what, we're really about making cars and we should do that. Or like, boy, the markets have turned. This is a really difficult time. I have to put my investment

here, or there's a new philosophy. Like actually we're not gonna do that, we're gonna do this other thing. And so, you know, when you, that company is, for this case, making a very rational decision. That's what they're about. And this other thing is,

Yeah, now they're doing electric cars. And so for us, I wanted in founding Aurora with Sterling and Drew, it's okay, this is what we're about, gang. And so invite people in who are aligned with that, have the ability to kind of set the direction for it, know what our North Star is,

And that allows us the freedom to navigate the changes. But presumably everyone's going to get to the same place eventually. This is sort of like moving from horse-drawn carriage. Everyone's going to have a car company that's going to look the same, either autonomous and electric. I think there's going to be certain tech... Like, yes, I think automated vehicles are going to be broadly deployed. But I think it's a hard enough problem that there's only going to be a few of us who can actually deliver it. You know, I take, for example, we all...

well, most people who've done high school physics kind of understand how flight works, but there's really only two aircraft manufacturers. And it's because it's really hard. And there's certain insights and certain kind of supply chain, all kinds of reasons why there's really just two, Boeing and Airbus that make commercial airlines.

And I expect in this space, there's going to be just a handful of us who really are able to understand the breadth of the technology, have the talent, have the technology, have the IP portfolio, have the financial support and backing to go make this happen. And so... Because it's very expensive. People don't realize that, you know, that $3 billion from Ford. And they're like, oh, no. I'm like, oh, no, that's not much. That, you know, it ended up not working out. Yeah.

Part of our business model is, yes, it's expensive early to invest and build the technology, but then our model is to provide the drivers a service. And so our business becomes very asset light at that point. And so if you are FedEx, you'll call up PACCAR and you'll say, I'd like to buy a Peterbilt 579 with the Aurora driver installed. You'll pay Peterbilt for the truck, and then you'll have a subscription to the Aurora driver, which will then drive the truck for you as it's moving packages through your business.

So you're launching it at the end of 2024 using a quarter between Dallas and Houston, a four-hour drive. Why there? We chose Dallas to Houston because first, it's one of the busiest freight corridors in the U.S., so it's an important place for our customers and partners.

You know, the weather is good. Now we've built the Aurora driver to be able to operate in different weather, but when we first launch it, we only wanna validate it. It's not gonna be snow today. So last time I was there two of the trips ago or three of the trips ago and it was like snow and sleet. Yeah, damn you climate change. Dallas snow is not great. Yeah.

But much less of it. And then, you know, the regulatory environment broadly across the U.S. is very supportive. But Texas has been very supportive, very pro this technology. And so we're excited to operate there. So meaning that they'll let you do whatever you want. Hence why I assume Elon's down there or the rest, everyone goes down there. I wouldn't go quite that far. Right. And again, one of the things that's worked really well for the U.S. is that it has had in general a very permissive

whether it's the way we release new vehicles, that is self-certification, or broadly, the legal system in the US, if it doesn't say you can't, then you can. And that has really enabled innovation. And so, whereas in Europe, what we see is in the automotive space, they have a thing called type approval, which means you have to build a car to this spec, otherwise you can't sell it. And the government has, somebody has to come and inspect and say, you're doing that. And that's the way it works.

That bounds how creative, how innovative you can be. And so I think this kind of American spirit, so to speak, is part of what's allowed the U.S. to excel in technology. And is this a safer straight line path? Is that why you want to do it there versus anywhere else? No, we just see it as a really, like, again, busy corridor. The Sunbelt, there's this kind of freight corridor across the southern U.S., which is key to logistics in the U.S. And so if we're going to pick somewhere to start,

You know, Dallas is kind of near the middle of that. It's a great place where we can operate and it'll allow us to pick up that corridor and support customers. And you have some employees there, some here and some in Pittsburgh. Yeah, our company is across the U.S. We have offices in Seattle, San Francisco, Mountain View, Bozeman, Montana, Michigan, Pittsburgh, Dallas, Texas, and...

So we talked about how few companies can crack autonomous vehicles, and this has been your challenge for a long time. You're doing a pretty straight path while Waymo's autonomous driving taxi services in Phoenix and San Francisco have to navigate pedestrians, bike riders, congestion. As a result, they won't be building them faster and smarter than a truck that's barreling down the highway at 70 miles per hour. They're not going to be doing that.

Yeah, I think that we have... So first, I would say that when we're driving on the freeway, it's disappointing. There are still pedestrians. I'll show you a video after. I actually saw one last night. Yeah, like we have this one where... I don't know if you remember the old video game Frogger. But we're driving on the freeway. There's lots of traffic moving at seven miles an hour. And this gentleman runs across the freeway between cars. It's insane. And so we have to be able to deal with that.

Recently, we were driving near one of our terminal sites and this shirtless guy was riding a bike down the side of the road in the wrong way. We don't get away from those problems, but we do think that the business will scale more rapidly because the freeways basically look the same, right? They've been engineered, designed to be the same whether you're in Texas. And they're federally. And they're federally regulated. So that means that there's a standard to them.

And so once we crack at working in one part of the country, we expect it to very rapidly expand. Because they're all the same. All the interstates look the same and feel the same. That's right. And then we have to, we of course can't kind of just get to the end of the freeway. We have to get off into the commercial neighborhood and get to the depot because it turns out that if you had to

you know, jump into the driving truck with a person to get it the last few miles, that wouldn't really be a product. So we do have to deal with some of the service street problems. Right. What most people don't realize is trucks don't usually come into cities that much. Smaller trucks do. Bigger trucks tend not to. That's right. There's a huge... Depot. There's a reason why these depots are around the edges of cities, because it's hard to drive a truck through a city. And so we expect to operate between a lot of those kind of terminals near the periphery. Over time, we'll ultimately take the groceries to...

the store uh but you know that's down the road there's going to be a store um you're expanding beyond texas where's your next state to use like not california where you have made your fortune here and you are located he made our fortune that's optimistic um we we'd love to operate in california um uh right now you know we're working you know the dmv here had uh there was a law passing that enabled automated vehicles the dmv created a lot rules for um for

Small vehicles, heavy trucks. They're much more in cars than trucks. And so at some point, you know, that'll shift. But in most of the U.S., we could operate today. For us, we'll be expanding along that Southern Freight Corridor. We'll be back in a minute.

So Tesla's attempts to get into self-driving mode have been well-documented. There's been a lot of issues over the years, and unfortunately, it's all been focused on Tesla because they're the only... When I walked in here, all your electric charging stations, which was a pleasure to see 20 or 30 of them, Tesla's mostly, but more and more, it was one of my cars. There was a Bolt there and others. Has that been problematic? Because obviously, they put these on the road before they were safe. Elon Musk has a different philosophy about safety.

Talk about your philosophy versus he wants just one or two signals. You believe in packing a car with signals. Packing a car with signals. Yeah, we think that to do this well, you want redundancy and that it turns out that that LIDAR, the radar, the camera, they all see the world differently, which means that they fail in different ways and they get you different information. And so if one of them doesn't work for some reason, it's likely the others will. And that brings increased safety, which we think is really important.

You know, we have really taken the approach of having trained operators in our vehicles as we're developing it. You know, these are folks that are part of our company that are bought into the mission. They're going to help us develop the technology and keep everyone safe. And we think that's the practical, safe way to do with it. It's hard for me to see how

you know, a single camera is actually going to solve this problem. I don't see it happening anytime soon. Which is on the Tesla. Which is what, yeah. And it's not to say it's not like brilliant, amazing technology, but there's a difference between amazing technology

and a safe viable product. Right, I'm not sure that's the concern, unfortunately. But you think it's done damage to autonomy because it was so like, it's gonna work in five years, three years, two years, whatever, and it doesn't, you know, and there's been one spokes model for this. I think it's caused confusion. And one of the things that I think is really,

is that if you wind the clock back six years, it was this technology is gonna be here next year. It's basically solved. Everyone's gonna have cars that drive themselves. And that was not true, but that was kind of broadly the zeitgeist, right? This problem solved and it's out there. And now the zeitgeist is, I don't know if this is ever gonna happen, but what's awesome as someone who cares about this, has been working on this for 20 years is it is happening.

That whether it is, you know, our friends at Cruise Waymo who are out there with vehicles on the road in San Francisco and Phoenix and with driving around and

serving the public and actually doing useful things. What we're doing with trucks, like we're pulling loads every day for customers and we're on the cusp of operating without people on board, right? It's actually happening. And the benefit to all of us, the businesses we're gonna build around this are gonna be great. Was it damaging to talk about it like it was happening tomorrow? You never did. I said some time ago, my mission was to try to get this to happen before my son got his driver's license. Didn't happen.

He still hasn't got his license. So I still technically got time. And if we lived in Phoenix, I would feel pretty good. So, you know, I feel like I'm off by a couple of years on that. But no, like I've had the fortune, misfortune, you know, I've worked in this for a long time and kind of understand the complexity of it. And we've been playing the long game and investing to go make it happen. But you need that energy, right? I don't want to fault people for optimism. I don't want to fault people for...

Well, it does set people back because everyone doesn't think it works now. Right. I don't. I thought it would, but now I don't. Yeah, but it will. And it is. Sure. I think will is different. And then, of course, everyone shifted towards electric as the goal for everybody. Let's get everyone in electric and then we'll worry about autonomous.

And I think that we need both, right? That electric vehicles or fuel cell vehicles, whatever you wanna use, the benefit for the environment and reducing carbon dioxide production, like that's immensely important.

But these are separate things. That have sort of been mashed together. Yeah, right. And it's just because it's kind of new automotive technology. Right. Talk about the unions. Has there been a lot of resistance to what you're doing from the unions? So we're starting to see some organized efforts by the Teamsters and others to push back on this. And, you know, they have a perspective.

We see the opportunity to improve safety. We expect to grow jobs over time with this. We expect to improve the economy, expect to make transportation more equitable. And so we'll continue pushing forward. So when you deal with the unions, Musk himself has had a lot of union issues. Does concentrating on software and smaller hardware mean the question of unions isn't as big?

For you, at least. I guess for us, we're really focused on how does this technology go out in the world and do something useful. And the trucking companies have come out and said that they expect their drivers to be there. They just can't get enough of them. And so I think this feels like it's kind of...

it feels like it's not quite accurate, right? The concerns that are being raised. But you see the unions doing this, the Teamsters and others. They absolutely are. What do you do? Do you call them? Do you do... So we tell our side of the story and we let the legislative process do its thing. Do you think it's a mistake on the part of unions to focus on job loss when there's not enough of them? I think that the concerns they have are laudable around safety and employment.

And I'm just not convinced this is really the right place to be pushing them because we are really pushing for safety and we do see an opportunity for job growth here. All right, speaking of job growth, you went public via SPAC in November, 2021. When you closed the deal, you said Aurora had $1.8 billion in cash and gross proceeds in November. The figure was $1.2 billion in cash. Is this very expensive? Yeah, because you're not building trucks. You're not building cars.

And you can't just rely on the stock doing its, you know, doing a Tesla thing for you all. And a leaked memo in September said that you need to raise another $300 million the next year, adding six months to your runway. So talk about that as the market cools, because these are cutting-edge technologies where people have to make investments in an economy that is not...

oriented towards that. So I've talked a little publicly about that memo, and this was a memo that I drafted kind of as the markets had been shifting. And it was really the responsible thing to do as any organization. It's like, okay, here's the world has changed. Let's make sure we look at the span of options from, you know, pressing the gas pedal harder to, you know,

should we sell the company? Had a really good board discussion and obviously we're still here and running along. And, you know, so you can infer what the outcome of that discussion was, but it like as a,

of our shareholders, of our employees. Like you have to have those difficult conversations. And I made the mistake of accidentally emailing that to everyone instead of just the board. And so, you know, that happens, right? But I was also really proud. Well, thank you for that. Yeah, you know, I get to do my part for employment. But, you know, I was actually also, after the moment of cursing myself for actually fat fingering this, read it again.

And it passed the test, right? You talk about internal communications, you want to be able to see it printed on the front page of the New York Times. I was like, you know, any of this. Fox News didn't get that memo, just so you know. The memo about the memo. But investments in autonomous vehicles is cool. You know, it was very hot. Now everyone's running towards AI. They're at crypto. That's just the way tech is in Silicon Valley. But there's a decline in investments in autonomous vehicles year over year. Are you worried about raising money? No.

Yeah, so we look at the landscape and of course we are thoughtful about how we're spending money. We've put effort into focusing our investments and building the team more cautiously than maybe some have. But when I look at the landscape around us, I look at what we're doing as a team, I've never been more enthusiastic or confident that...

you know, in the space of trucking, most of our competitors have fallen by the wayside. Yeah, just Argo AI, you said, was backed by Ford and VW pulled their funding. Otto, which was bought by Uber in 2016 and then shuttered in 2018 in favor of their Uber Freight project, which is a partner of yours. Is it just the culling of the herd here, so to speak? Yeah, I think that's what's happening, right? And that will allow the capital markets to understand what, you know, kind of

who's gonna be a winner in this space. And we are keeping our heads down and executing. And the conversation I have with investors, they get it, right? They see the opportunity. Trade transportation in the US is about a $700 billion business. So we're playing in a space that's what, four or five times bigger than what Google. Someone has to win. So there's gonna be somebody winning the space, yeah. Now, who's your main competition, Waymo? They also have a trucking division. Yeah, so as we look at the market,

We see folks like TuSimple is one of the companies that's been out there that's been dealing with some challenges.

Embark was one that shut her down. And then Waymo has a trucking group. And, you know, I had the privilege to help build that team. And it's an amazing team. But they seem to be more focused on passenger vehicles today. Which you are also. You have some. In the long term. It's kind of, we're almost the inverse of one another. So I want to finish up talking about regulation. Tesla moved its engineering headquarters to Texas in 2021 before Elon Musk.

announced he was moving them back to California just a year later. But he's building a town in Texas. You're launching Aurora there. Talk about what regulation you need that you think is important. Yeah. So in the near term, we don't really need any regulation, right? That again,

If we were confident in the safety of the vehicle, we could launch it today in most states. What we would love to see is at a federal level kind of guidelines, rules that cover the whole U.S. Because it turns out there's this thing called preemption, which means if the federal government has a law for safety in these vehicles, then that supersedes whatever's in the states. And so one set of rules across the whole country would be preemption.

Where is that right now? So there's been a couple of different runs at federal legislation. We'll see. They're still working on privacy legislation around the internet. I don't know if you know about that. I've heard a little bit about it. They're going to pass it someday, sure. So at some point, maybe this will move. Yeah. We'll see what happens. And do you need it? Does it have to happen? No, we don't need it. It would just help the U.S. be more competitive.

With who? With China, with the rest of the world. Who is the biggest international competitor? Obviously, China's very big in autonomous and electric cars. I think China is really the biggest competitor. And why is that?

I think one, there's a lot of smart people there. And then two, the government has really put a lot of energy behind it. They've made this one of the priorities for them was delivering automated vehicles and they've created both a regulatory and financial environment where they've enabled that. And so, you know, I don't see

Chinese players really being able to operate in the US and I don't see US players being able to operate in China. But if we play our cards right as a country, we can be the international source for this technology and again, continue to- And they're putting a lot of focus on it. Of course, yes. National focus. Yes, they are.

Do you feel our government has not done it? They haven't done it in space lately. They haven't done it in AI. You know, I just interviewed Dan Altman. He said, I wish the government was more involved. Yeah. I think it's been a tough few years. I look at back when I was a grad student. Decades, really. Right. The DARPA Grand Challenges. And this was a few million dollars by a defense research program that really...

energized, kickstarted what has become the self-driving vehicle industry. Yeah, you were all from there. Right. And it was just kind of the poster child of American investment by the government done really well. And we haven't seen a lot of that over the last few years, right? Somehow we're kind of missing that critical step. And so I'm a little worried broadly about policy, whether it's investment in base technologies or whether it's

you know, immigration that's restricting the flow of incredibly talented, educated people. You know, that's part of what's made America competitive for decades. And when you think about a truck, you know, you were saying we were in a cab, we were in a big cab. I was pretending I was a truck driver. But these cars won't have cabs, right? They will look different. They will not look like what trucks do now, except for the back part. Well, I think that the

For the foreseeable future, they will actually. So today we use trucks that have a sleeper cab at the back. And you put the computer back there. We put the computer back there. We have some room for our engineers to ride along with our operators. I think over time, that'll get, we'll kind of,

remove the sleeper cab, that'll make the vehicle able to turn tighter. It'll make it less expensive to produce. There'll be some changes to the aerodynamics, but it's going to be valuable for people to be able to get into these vehicles, to move them around and drive them, right? Whether it's their service. So I expect that that'll be there for a long time. Ultimately, maybe not. Maybe there is a new form factor that happens, but it's, you know, we're

We think about... Hyperloop, perhaps? No. Yeah, we'll drive on the roads that are out there. Yeah, that have, they exist. Yeah, good luck with those. So, last thing I'd like to ask, there's a lot of conversation right now about the need to regulate AI as it grows, obviously. I think, you know, Death Flow to Humanity was the most recent memo. With autonomous driving, which uses AI, what regulation do we need to grow responsibly? As someone, as a leader in tech, you obviously have to use AI as part of your...

No, and we've been using really like the- Or maybe say machine learning. Machine learning and AI, right? The old joke was that AI was a part of machine learning that didn't work or machine learning was part of AI that did work, whatever. Little geek humor, thank you for that. Sorry, couldn't let you get away without that. But we've been using kind of the core parts of what's really fueled this AI revolution for the last six years. It's been kind of part of the underpinning of how we predict other vehicles are gonna behave on the road.

And so, yes, we see that as essential. Obviously, our applications are much narrower. So where could we see some help with regulation? Again, it's around, we think we're doing this about as well as you can on terms of safety and thoughtfulness. And so at some point, kind of,

putting a floor under what is acceptable, I think will be meaningful, right? And it will help protect the public. I think one of the things that people miss though, is that even today, there's enough tools out there to protect the public, right? That the police can pull vehicles off the road if they think they're operating unsafely. The DMV, the federal government can recall vehicles they think are operating unsafely. And so,

You know, it's really about how do we level the playing field? How do we accelerate the technology over time is where we see the opportunity with regulation. Well, Chris, I'm so glad I didn't run you over those many years ago. A little bit. Me too. Really appreciate it. Just a tap. That's all I wanted. Comedic timing. Yeah. Anyway, I appreciate it, Chris. And it was great talking to you. Thank you. Yeah. Thanks so much, Kara. Appreciate it.

I most enjoyed his origin story for getting into cars. Yeah. I love that he got into it because it was just cool. That's the story of a lot of people. I mean, you know, a lot of people have this dream of, you

you know, two things, jetpacks and driverless cars, right? You know, that kind of thing. And a lot of sci-fi. There's so much sci-fi around it that it's crazy. Like there's tons and tons of sci-fi. But the idea of driverless cars is a big one. It sticks in the human imagination. So it's- It is so cool to see these technologies. I used to go to the Hyperloop competitions that Elon had in Los Angeles and I loved it. It was like cool runnings with a lot of Scandinavian high school nerds.

who would come and build these extremely fast vessels to go through the loop. It is an industry that kind of attracts, it's kind of like sports in that way. It attracts a certain passion builder, and he certainly was that. But he versus Tony Fadell, who we had on as a guest, seemed to have very different experiences of Google or, I guess,

Chris was more diplomatic. Yeah. I mean, I think, you know, I think he was frustrated too. I think he was, he wanted to do this and it was just one of Google's many plots and schemes. You know, that's the problem with Google. And it's kind of great that they have this sort of these units to do. And it's been pulled back quite substantively in recent years, but you know, they were never going to be the leader here. They're just sort of seeding, trying to seed the ground to make it. The problem of Google is that they're an advertising company that's like subsidizing

small forays into other markets. Yeah, they did other things before advertising. So I think it was great that they were doing that. They did that around Android. They do that around a lot of things and wacky things too. At one point, the Google founders wanted to build chairlifts in San Francisco. You know, I don't mind. I don't mind. They're crazy. It's their money. It's shareholder money, but it's really their money because they control the company. I am worried about things like the California Assembly Bill, which is requiring human drivers to be present in

You know, the Teamsters are lobbying for this. Why is having a human in the car such a downside if it's just a bridge to getting to autonomous? It allows a testing, right? And it allows a fail-safe or a backup while people get used to it. Yeah, that sounds good. There's no humans at bank machines. This was another thing with tellers. Like, it's just like you don't need them eventually. You don't need them. And that's something...

Travis Kalanick said to me in a much cruder way when I did an interview with him about this, when you get rid of the guy in the front seat, it's a really good business. And eventually, I'm sorry to tell you, business is going to win.

business is going to win over all of this. It always does. Yeah. The question is, how do you engage people? How do you allow people to have some quality of life? The other day I saw this post that AI won't replace you. A human using AI will replace you. It stung more in some way because I'd rather be replaced by a computer than a human that's better at using computers. Well, sorry to tell you that's what's happening. I mean, it's just job preservation is going to be the big topic, whether it's AI or autonomous cars. And there will be other jobs. This has happened

historically has happened again and again and again. And, um, and trucks driven by humans, as much as I love a trucker movie, I love Burt Reynolds. I love convoy. I love all these things. Um,

We should have drivers there until it's safe not to. And safety should be the prime thing. But it is inevitable and there's a lot of jobs that humans should not be doing and just do it just because that's the way it's done. Yeah. Safety of the individual cars and then safety of the overall system, the cyber system, the hardware, the software. Yeah. It was funny when you kept asking him the question of if Elon was helping him or hurting him.

in these predictions. And he kept being very diplomatic, but you had a point of view that Elon was unhelpful from. Yeah. I think it's unhelpful to make people think it's going to work and not get used to the losses that are going to need to happen and the problems. And over-promising and under-delivering is always never a good thing. For the industry or for... Yeah. It's good to make people dream of what we could do. I feel Elon has actually done that in huge ways. I mean, even that Hyperloop competition I was making, he brings imagination to reality.

I love Tesla Elon. Not everything about Tesla Elon I like, but I like that idea. And I think it's a great idea. But at some point when you actually don't deliver and then look disingenuous, it's a problem. That's all. Yeah. Or it's a white space for people like Aurora and Waymo, where you set those targets and not met them. It creates a space for them to do better. But let's see. Right. And it'll be an issue for states. It'll be issues for cities and

how robo-taxis operate, how autonomous cars use, if they're going to ban cars from the center of cities, things like that and making it a better city. But the cities have much more problems on getting people to come downtown at all. So it might not be a problem not having... I was in San Francisco and I was able to park the other day easily in downtown. Well, I keep wondering when you get, when Aurora sends you this,

you know, this Aurora truck just for a trial. Where are you going to park it in D.C.? Your neighbors are going to like that. Next to my electric Hummer. Well, these are big problems, where to park your electric Hummer and your Aurora truck. But other people might have problems too. Questions, want advice, and we are doing a special advice episode of On. And Kara, they can follow us, right? Yes, it's 1-888-KARA-PLZ. And, you know, ask me anything. Anything? No. Well, I'll ask you to read the credits now if you don't mind. ♪

Today's show was produced by Naeem Araza, Blake Nishik, Christian Castro-Rossell, and Megan Burney. Special thanks to Kate Gallagher and Brad Chun.

Our engineers are Fernando Arruda and Rick Kwan. Our theme music is by Trackademics. If you're already following the show, you get a car, you get a car, you get a car. No, nobody gets cars here. We're not Oprah. But if not, I will come collect your car in a tow truck driven by yours truly, autonomously, of course. Go wherever you listen to your podcast, search for On with Kara Swisher and hit follow. Thanks for listening to On with Kara Swisher from New York Magazine, the Vox Media Podcast Network, and us. We'll be back on Thursday with more.