The Road to Autonomy: Chris Urmson on Building Aurora

I want to build a business that's self-sustaining. Freight is the right place to do that. You know, our customers at Aurora are carriers of FedEx's, Werners, Hirschbox, Schneiders of the world, plus the trucking industry at large. 90% turnover year. So imagine trying to build a business where nine out of 10 of your employees aren't here next year. That's just incredibly challenging. We just slot in and fill in for some of these folks who are leaving the market. In 2009, I got recruited to go to Google to help found a secret project which is now Waymo. The first time I got on Waymo and I was just like this is awesome. I feel like incredibly lucky that I get to work on a technology that is just super interesting, hard and ultimately will be incredibly valuable economically as well.

Oh, cool. Yeah, I've been really looking forward to this. I was watching a clip of Elon this morning. From his was it maybe his investor day or something? Where he was talking about FSD and just the progress that they're making. And I just it made me even more excited to talk to you about all of this. How do you think that progress is?

I think that if you go to San Francisco and ride in a Waymo, you're there, right? I think, with a little bit of luck, a month from now, you get in one of our trucks, driving between Dallas and Houston, it's driving itself completely. It'll do so safely. It'll be useful for folks. I think that's that's a real thing. I believe and our team believes that you know to do this in a way that's safe and robust you want to use a variety of different sensors because each sensor has strengths and weaknesses and you can fool different sensors you know not even intentionally just by looking at the world. And so by using multiple sensors you get to the point where you can be really robust and reliable. And without that, it's kind of hard to imagine doing it in the in the near term.

No, I did not. I drove I drove my Rivian here. But I'm I'm thrilled, right? Like that, it must be special for you. It really is, right? It was it was hard to leave. And the folks that stuck around and kind of continue to build that over the last eight years. It's just awesome, right? It's doing it's doing the thing that we expected and hoped. All right. I remember talking with Dmitri, one of their co-CEOs, when we were working on this together, and it's like, if we get this right, this is going to be better than an Uber, right? It's going to be private. It's going to be safe. It's going to be clean. It's going to be, you know, and people are going to ride in our thing and they're going to say, 'I'm never going to take an Uber again, right?' And and I love Uber. Uber's a wonderful platform, but this the the opportunity with this technology is profound. And to see it actually happening and you know, in San Francisco, it's just at this point, it's just part of the city. Right, it's and that's I think when something is real, when it's just there. It's kind of magical.

Yeah, it was unbelievable. It's it's real and it's happening, right? And so when you ask a couple minutes ago like when's this happening? Like right now. It's awesome.

Yeah. They're very similar, right? In that we as far as I understand, both of us use a combination of laser, radar, and camera to see the world around us. We use high performance computing. We use AI and machine learning and all the other good algorithmic stuff to make the thing understand the world around it and drive safely through it. And so it is yeah at that level very similar. At Aurora we've got a couple of really interesting tweaks on that. So we have a special kind of laser rangefinder that allows us to see further than conventional LiDAR can. And that's really important for us because we're driving big trucks on the freeway at highway speeds. Like semis. Yeah. 18 wheelers. On the road, 70 mph down the freeway. So, you know, that that kind of sensor I think is is unique to Well, I don't think I know it's unique to our approach to it. And then of course we're working on big 18 wheelers and they're working on passenger cars at this point.

It it's phenomenal. It's and you know for our company you know when we founded the company you know it was like well we can kind of figure out within a couple years when we're going to be able to launch and then it got to within you know Airbar was a year and then it was quarters and now we're in weeks. Right. in our certainty of exactly what will be on the road.

They're saying they only need camera. They only need camera. Yeah. And but like they're making a lot of progress. Oh, and and technologically awesome, right? Like, but do you think there's something limited there? Yeah. I I think that they're not going to be able to cross the chasm, right? And there's this fundamental difference. Sorry, what is the chasm? Yeah, that's where I was going to get to, right? There's this fundamental difference between helping a driver where the driver is attentive and paying attention versus replacing the driver in the vehicle. And why do I say that? So, let's may take Tesla off the table and let's just talk about, adaptive cruise control. So, this is the thing where you're driving down the freeway, you set the cruise control. My car has that. Yeah. It slows down and speeds up based on the vehicle in front of you. So, the assumption with adaptive cruise control is that you're looking out the window, you've got your hands and feet ready to go. And so it's actually okay if it doesn't hit the brakes. It should hit the brakes. But it's okay if it doesn't because you're paying attention. And even if it gets it wrong half the time where it needed to hit the brakes really hard and it shouldn't have, you're there as a backs stop and you're paying attention. So, you're going to take over. And what that allows it to do is to not have false positives because if you're driving, I don't know what you drive, but you know, you're driving your nice BMW or Mercedes with adaptive cruise control down the freeway at 60 mph or 65 and it hammers the brakes on in the middle of nowhere. You're like, 'That's not cool.' You do it a second time or a third time, not only you creating risk for other users on the road, but you're now calling the car dealership and saying, 'I want my money back. this thing's a lemon. I just drive down the free freeway and hit the brakes whenever it wants.' And so when you get into, you know, the machine learning space, the AI space, you're willing to accept a lack of recall, which means kind of seeing the seeing the events, so long as the precision, so the events that you do see are real events because that's what the customer cares about. And the kind of the product requirements are react when you can, but whatever you do, don't have false positives. Don't hit the brakes when you're not supposed to. And and so that you have a freedom on the on the recall dimension. When you start talking about a self-driving vehicle, one where it's taking over the whole driving task and you're not expecting someone to pay attention, you no longer have that freedom because if you know, if it crashes into half the vehicles he's on the road, that's just not a product, right? Whereas in the the driver assistance case, if it didn't break for half the times it should have, it's kind of okay because you're going to take over and hit the brakes.

Not possible is a very strong statement but I don't see how you get there technical like yeah and and the again part of it is market forces right that if I'm making a driver assistance system I want to find this operating point that allows where the the false positive rate the false breaking events are so low that I can accept it and that it behaves well enough that the customer wants it and then I'm going to take all the cost out of that hypo possibly can. Where and and so that's going to push you in a direction of down costing because you've you've kind of met the goal to go from that to a self-driving system where you have to have both high recall and high precision. You're using a different set of technologies to enable that. So it's kind of how I think about it at least.

Well, no, I didn't write in the beginning. I I absolutely did not write all the code, right? Like we had we had an amazing team who did the work. You were driving the technology strategy from the ground up writing early lines of code. Oh, absolutely. For that system. Yeah. No, I've been I've had the the good fortune to work in the space for 20ome years at this point and you know back in the day working on robots to race in the DARPA Grand Challenge and then helping found what's now Waymo and now building Aurora. And yeah, I get I had the privilege of working some of this technology, but I I I definitely did not write most of the code.

I think so. Yes. Yeah. Like you're I like to think so. You're a you're an engineer turned CEO. That's right. Yeah. And and in the DARPA challenge that's how you got to Google in the first place isn't it? Yeah. Well, I had been doing so there's DARPA challenges. These were the idea was back in 2003. The defense department wanted to accelerate automated technology for the military. And so they had a competition to have vehicles drive from Los Angeles to Las Vegas across the desert. And so I was at Carnegie Melon at the time. I was a grad student and became the technical director for Carnegie Melon's team in that. a lot of good stories there. But long and short of is it it took about three years and we got to the point where we could mostly drive on our side of the road, mostly not crash into things. And then the defense department says, you know, problem solved. We'll leave it to industry to industrialize it. And then in 2009 I got recruited to go to Google to help found a secret project which is now Waymo with the hope that we could take the ideas that we'd come up with at the DARPA Challenge and commercialize it and build self-driving cars for folks.

Yeah. So, so this was a competition again. They they put out a call to anyone who wanted to come play and they said if you can drive across the desert from Los Angeles to Las Vegas and you can do that in less than 10 hours. The first team that does that on race day is going to get a million dollar check which at the time sounded like a lot of money. Still a lot of money. And so bunch of us grad students, couple faculty members, we worked on it. The first year we took this Humvey and we cut the top off it, put a big electronics box that had a computers. I don't know if you remember itanium computers, but we had one of the first itanium computers in it. We had that that actually I think of as kind of the Lucy of self-driving cars, you know, the missing link where it was the first time we had lasers, radar, camera, high performance computing, and mapping, high definition mapping together to get a system to work. That first year, we're supposed to drive about 150 miles that trip. We ended up going about seven and a half miles before literally bursting into flame. Well, almost literally bursting into flames. So it was a little bit disappointing. It was honestly it was kind of crushing, right? You've worked incredibly intensely for nine months. But the defense department and in retrospect we recognized that we had basically driven, if you looked at the product of distance and speed, we were an order of magnitude ahead of what anyone had done before. despite not meeting the goal. And so they said, you know, come back next year. We'll give a $2 million check to the winner. And so we came back with Carnegie Melon. Stanford competed that year. And at Carnegie Melon, we ended up coming second and third behind the Stanford team.

This was 2005. Okay. And so that was exciting. And you know, it was those first two events were kind of like robot nerd Woodstock, right? It was a bunch of folks out in the desert and hanging out and you know it was great camaraderie. And so then the defense department said okay in 2007 we're going to ask you to bring vehicles and at this time we're going to close down well they didn't say this but what they did was they closed down an air base and we had to drive around with moving traffic. So they hired a bunch of stunt drivers out of Hollywood got them in cars brought them out there and so we've got robots driving around this old air base. And they had to drive 60 miles around the air base, you know, going various places with interacting with traffic and whatnot. And that one, the Carne Melon team I was the technical director for ended up winning that one. The Stanford guys, I think, came second. So, and that that nucleus actually that started what's now Waymo, what was called Project Chauffeur when we kicked it off, was a couple of us from the Carnegie Melon team and a number of folks out of Stanford that kind of built it in 2007.

Yeah. So, it was driving it speeds up to, I don't know, 35 40 mph. it was driving on its side of the road, can make left and right turns at intersections, it could track other vehicles, it could take turns at stop signs. So, it was kind of the the toy problem sounds dismissive. It's not the way I intend it, but kind of the the proof of life that you could actually go do this for real on the real world.

I don't know that was all the way there yet. No, I thought like I thought this was cool. It was clear how it was going to help the military. It seemed likely this would be useful. But I was still like it I I I'd taken a brief stint working for a defense contractor. it was faculty member and it seemed like it could be but it was not clear that it would be. Yeah. It took a few years and some of the progress we made at at Google to get to the point where like no this is going to happen.

It was it's it's been technology I think for many years it's oh I don't know if regulators going to be okay. I don't know. And it's and that felt always like not it right. That the society I'm I'm I guess I'm a bit of an optimist. You kind of have to be to be an entrepreneur, I think. But especially to be one in self-driving. Geez. Yeah, fair enough. But that I I believe that society when something is valuable, we find a way to use it, right? And and make space for it. And it's clear the technology that we're working on, you know, has that opportunity, right? We can you save lives. The US alone, 42,000 people die on roads every year. If you contrast that with air traffic where up until very recently, it had been 10 years since we'd had a fatality in American airspace with an American flag carrier. you know the the need right if you're if you think about personal mobility that as people get older they lose the privilege of driving and that is an accelerated death sentence. People are social animals and you know you take the keys away and you control for all the other factors you die much more rapidly when you don't have that ability to be out and be part of your community than you do if you continue to have that freedom and ability to move around. in the space that we're focused on with the Aurora driver in trucking and freight that over the next decade we need a million new truck drivers to support the the both the number of folks leaving the industry but also the growth of demand in logist in the logistics space and there aren't a million more Americans who want to go do this and with the evolution of our immigration policy it's unlikely we're going to be able to find other people to come into the country to do that job. And then if you're a company that's going to use this technology, you know, our customers at Aurora are carriers. So, you think of them as trucking companies, the FedEx's, Werners, Hirschboxes, Schneiders of the world. They have a hard time hiring drivers. So, across the trucking industry at large, 90% turnover year. So, imagine trying to build a business where nine out of 10 of your employees aren't here next year. like that's just incredibly challenging. that because of really good safety reasons, you don't allow the people to operate those trucks for more than 11 hours a day. And so you have this expensive asset that you can operate less than half of the day. And so if you think about what the Aurora driver will provide these companies is it'll complement the people they can hire to drive. it'll allow them to run the asset twice as hard and that's a key financial metric for them. And it'll give them a scalable access to a safe driver. It'll ultimately drive down injuries and fatalities on our roadways. It'll reduce for them the cost of operation. It's sustainably a huge thing. So today our fleet is seeing a 13% improvement in fuel economy relative to kind of the the national standard. And so, you know, it's I feel like incredibly lucky that I get to work on a technology that is just super interesting, hard, and is really important, and ultimately will be incredibly valuable economically as well.

Kind kind of incredible, right? I I can't thank them enough. You couldn't believe that call. I mean, no, they who one of them called you? No, they did not. Sebastian Thrun, who is led the Stanford team that we competed with. We had been friends. He had come to Google to help build Street View. So if you think of Street View, that was kind of him and his team delivered that, which is kind of amazing product in of itself. And we had been talking about maybe starting a company to build self-driving technology. Glad we didn't at that time. At this time, it's a good time. Then it, you know, it was just too early, it turns out. But Larry, Larry, I think in particular, was like, why would you do that? Why not just do it at Google? And I, you know, I think there was a, you know, Larry can speak for himself, but my my understanding, my belief was that Larry looked at the resources they they had at Google, the incredible people, technology, like this is important what we're doing. We can do more. and saw, you know, transportation as an opportunity that really needed to be addressed in terms of the safety impacts, the you know, the accessibility, the economics of it. And was like, hey, well, we should do this here and start it. And so when I joined in 2009, February 2009, it was a black project. Like not even within Google they knew about it.

Totally. I I am sure all the counting was fully above books and all that. But what I can say is that most people at Google didn't know, right? There was a very small why I think because we he wanted to protect it, right? Did not want protected from too many people being excited and interested and wanting to get their hands. I think that at any large company there's a set of cultures and kind of bureaucracy and stuff that happens for good reason and it was like okay this needs to get incubated get a chance to kind of grow and mature. And so I think it was a let's just create a pocket to allow that to happen. And so and they did right to very much to to Larry Sergey and you know Bill Coughran their credit and so we kind of it it was about two years. It was just a black project. I remember we talked about it ultimately at one of the TGIF events at Google and everybody kind of lost their mind. It was very cool.

Yeah, we were so at first we shared space with some of the geo team their their mapping team but ultimately we ended up having you know kind of a different building and you know we were kind of we were around the Google venture people a little bit for a while and then and if people asked you what you do here it didn't come up a lot you know it just kind of work on go stuff right you know and yeah I remember even being out in the public with these things and we had, you know, the big spinning lighter on the top. Looked like a Kentucky Fried Chicken bucket. And, you know, we had this thing called a wheel encoder. Measured the rotation of the wheel. It had a stick on up. And I remember at one point being on 280 filling up a gas station and somebody comes up and it's like there were two thesis. One was that it was a perpetual motion machine, right? That this thing on the top was a wind turbine and we were carrying picking up energy from the wheel. It's like cool. you think that that's why we're at the gas station. And then the other was that we were stormchasers, right? That that we were out there. And so, you know, we just kind of nodded along and, you know, didn't actively mislead, but also just kind of kept to ourselves and did our thing.

Because because of funding. I think that it the technology wasn't ready yet and it takes a place like Google to have the patience for you to be in there taking your time. Whereas if you go raise money and you do all these things, the time horizons wouldn't really work. We certainly weren't taking our time. Like this was a pretty intense this wasn't kind of a sleepy research lab. like you know I was getting up and head to the office at 8 or 9 you know work till 6:00 go home have dinner put the kids to bed go back work you know 8 9 till 1 in the morning right and there was a core of us that were doing that for a couple years right so we were we were really working hard at it right we wanted to make it happen. but the investment that was needed to kind of just understand this right and incubate it that you know it needed the resources of of a Google and and I I I sincerely believe that without that without the urban DARPA grand challenges none of this would happen and then furthermore without Google's willingness to invest and kind of explore what seemed like a crazy idea at the time you know we wouldn't be where we are today there certainly wouldn't have been you know a cycle of investment that we saw over the last call it you know 8 to 10 years in the automated vehicle space at all.

I I think for the first like we were pretty lean and you know without trying to reveal internal Google expenses but like you know over the first two years we were probably got up to about 30 to 50 people. Okay. started with kind of a half dozen. So, you know, it wouldn't have been that much money to raise, but I think the ability to raise, hey, I'm gonna go build a self-driving car company, which by the way is not a word that exists today because we made that up when we were at Google. And and I don't know exactly when it's going to work and you know, like I don't clear how the business model is going to like Uber didn't exist, right? So even the notion of kind of ride healing and you know taxis were the biggest metaphor but how big is that mark? I just don't think it could have happened. you know the amount would have been modest you know 20 $30 million by today's standard but you know if you look at how long it's taken Waymo and again this is not lazy people they're working hard and focused on it. you know, they started in 2009 and it's 2025. So, 16 years to get something that's starting to look like scale and real. That's a long time to That is a long time. Yeah. And and when 16 years just to get the first cars on the road. Yeah. 15 years for the first cars, but Yeah. Yeah. Exactly. And this so this is this is why I'm saying and I feel at Aurora you know we've been at it eight years now and you know we've been fortunate to have investors who understood the mission we were on and we've we've worked hard to make sure they understand this is not a three months in hockey stick this is an eight years and you know transform an industries play and so I'm incredibly thankful for their patience but 16 years is even more crazy Yeah.

Totally. And what we think of as AI today was not what it was then. Absolutely not. and and and so I I think there's a there's an expression ignorance is bliss and I think that was key to this, right? And the kind of the more engineering and business one is, you know, think like an amateur or dream like an amateur and execute like a professional. And I think if we certainly if I had understood how hard this was going to be, I think I would have found something else to do.

I don't think any of us did. I think I think certainly in 2009, you know, it was like this was hard, but how hard can it be, right? And and I think that's key to any endeavor. I think if you ask anybody who started a company, like truly you knew how hard this is going to be, maybe for second or third time founders, but anybody who starts the first time does not appreciate what they're getting into.

It's always been people problems. right. That the technology I again maybe too optimistic. The technology you're going to figure it out, right? You you put enough sweat into it and you get smart people around. We'll get that figured out. the people stuff is always delicate and at least for me the stuff that's the hardest, right? not be because I care, right? I I want our people to be successful. I want folks who work with me, if they're investing their time and their effort into it and they're they're kind of gracing this with themselves, I want that to be worth it for them. And so it's always been people things like the, you know, there's been a lot of press around the Anthony Lewandowski Google stuff that happened back in the day and that was very stressful and difficult to navigate. leaving Google was a really tough decision. because again that company had been incredibly generous to me and given me a privileged and opportunity to do something amazing and ultimately it was time to get out of the way.

I was deposed and Google settled the day before Larry had to testify and so I didn't have to do that and but I did have to testify in the what do they call the arbitration dispute that he had with Google. So yeah.

There were a couple moments where it felt like I remember and and I think this all came out in the This is all out there. I'm I'm pretty sure, but I I remember so 2011 we had hit a milestone and Anthony was like, 'Hey, we should just leave.' and go do this on the outside. And I was like, 'No, I don't think that's right. I think, you know, Google's invested with us. We're here. We kind of owe it to them to keep building this. At least I feel like I owe it to them to keep building this.' And I will never forget that a couple of the senior engineers on the project at the time, you know, it was we'd been working late kind of, they were like, 'Hey, Chris, can we go talk?' And they take me into a conference room. the three of them and me and they're like, 'So, we've negotiated with Anthony. We want you to come be CEO of this thing. you know, we're going to give you the equity that you should have in it and let's go do this.' And I'm like, guys, you know, so it was one, it was very touching, right? Like these are these are people I continue to have a ton of respect for and really, you know, they're very great folks. but I like no, you know, and and it was very it felt at the time there was a real risk that that yeah, that that would tear it apart, right? These were these were folks that were key to what we were doing. They were again people I have a ton of respect for and really admire and so if they end up you know kind of falling under the pi piper spell and kind of get pulled out then to Uber well at the time it wasn't Uber this was to go start this was this was before Uber Uber wasn't until 2015 16 this was 2011 to go start something independent I was like okay well that that's kind of the heart of the team gone.

There was probably, I don't know, call it 40 engineers. 40. Yeah. And you're losing like four or five key people just like that. Well, more except these were four or five where this was three who came and they recruited a few others. Well, well, they didn't. So, they ended up not leaving at all either. Right. They they ended up only Anthony did. Yeah. He didn't even leave at that time. Wow. and again, go back to my inability to kind of communicate the strife that was happening. The strife. Yeah. So, yeah.

It it it never fully fell apart. That's the magic, right? That that if it had Waymo wouldn't exist today. And so those core engineers, those core people stayed. Yeah, they did. some of us left in 2016 ultimately. But I think by that time as you hope for anything you take, you know, you build like it was able to outlive you, right? And I think for me that's one of the things that's amazing like I I've take, you know, what Waymo's done since I left is due to Dmitri the Teendra the team that's done that work and you know it's awesome right and it's it's a lot of hard work and they they had to wrestle with a lot of things but there's a part of me that will always be really proud of the fact that I helped kind of set the set the culture set the organization you know set the path that got that to a starting point and then they you know carried it forward and it's awesome I I will always be thankful to Dmitri. You know, at the time when I left, you know, I was pretty disappointed that he didn't come with me and we didn't go do something together, but in retrospect, I'm really appreciative of the fact that he did stay and that he has been such an important part of driving that forward. And it's meant that that time that I spent there, you know, is resulting in something meaningful in the world, right? And that's cool. That's really cool.

I've been sell a lot of self-driving cars. It was just awesome. Actually, the first time I got away, I went for a ride with Dmitri. We went around San Francisco and it was just like this is awesome, right? And it it's doing the thing. you know, it's kind of like when you're I can't even think of the right metaphor with the kids. probably like when so our kids both are off at college and you know you you get them off to college and they're like oh they're kind of like a full-on real human being grown up. right and it was kind of you know similar kind of pride.

And and I think I think like anything else, certainly anything that's transformational, people will naturally and appropriately have skepticism. right, that you know, you think of like elevators. for a long time, you had to have a person riding in the elevator to press the buttons. because you know it was kind of scary otherwise right and you know when cars automobiles were first being sold in the US in some states you had to have a person walk with flags in front of the car so that it wouldn't scare the horses right like there's just like things that happen where it takes time for society to adapt a little bit. I think there's other times where the value is so profound that even though you couldn't have imagined it a moment ago, it happens pretty quickly and and you think of like the iPhone or Uber or Airbnb some of these. Yeah. Yeah.

I'm a little still, you know, my experiences with Airbnb have been a little mixed. But Uber certainly, you know, the iPhone I think it was like if you'd asked me in again 2006, do I want like a piece of glass with no phone digits in my pocket? The answer it's like what for? Whereas in 2011 it was clear that you needed these things. And so I think that my experience with people riding in self-driving vehicles has been that independent of the initial state, it goes through like a very short sequence, right? So, at first either they're super skeptical and they're like, you know, this is BS. Let me look for, you know, the dude hiding on the seat driving or whatever. Or it's terrifying or this is amazing, right? Those are kind of the three states they start. They get in the vehicle. For the first, you know, two or three minutes, it's oh my gosh, you know, there's no one touching the steering wheel and it's driving and there's no one touching it. And that takes two or three minutes. And then by about, you know, 5, 10 minutes, they're kind of like, is that all it does? Like, yeah, we're doing 65 miles an hour in the 101. Yeah, that's all it does. It's a self-driving car. What you expect? And then by 15 minutes they're on their phone, right? And like, yeah, and it just happens, right? And so, you know, your cousin's from Vancouver, you know, get them to get them one. And I bet you that's the way it plays out. And my theory on this and why I think it normalizes pretty quickly is that we're all used to being driven, right? Whether it's riding a taxi or an Uber or your significant other driving or your parents driving you as a kid, right? It's just kind of normal to be in a vehicle and have somebody else driving it. And so once you get over the there's not a person sat in that seat doing it, the experience is the same. And then similarly, when you're walking down a street in San Francisco, all those other cars you're not controlling and they're driving by. And so, yeah, at first it's kind of crazy that there's nobody in that, but very quickly it's, oh yeah, it's just another vehicle driving by that I don't control.

I'd love it. Yeah, be better than mine.

I think that's right. I think it's part of the human, you know, part of what makes us human is that adaptability, right? And you, you know, you think about the folks fighting in Ukraine today, right? Inhumane, awful, and yet that's just there's folks who are able to adapt to that, survive it, work through it. So yeah, I think that's also true. Yeah. Yeah.

No. No. I believe when you work somewhere and you disagree with what's happening, you've got kind of three choices, right? You can help fix it. You can get in line or you can get out of the way. And there were some decisions that were being made that I didn't agree with that I tried to address, tried to steer in the way that I thought would be most effective for the company to be successful. Was unsuccessful in doing that. Didn't really believe in the trajectory that we were on in the moment for whatever reasons. And given that the team reported to me and it wasn't fair to them and it wasn't fair to Google that I be there. And so it was time to get out of the way, take options.

It was more people and some strategy things.

Totally. I, you know, if you'd asked me even a year earlier, I'd have been I'm going to be here for the next decade building this. Right. I, you know, it turns out I have just somewhere else. You know, I felt incredibly like just incredibly lucky, right? That I had somehow stumbled into this opportunity to build something transformational to do it at a company where I believed in the values that we espoused and got to work with amazing people and it was hard to imagine something better.

Yeah, I'd worked for a defense contract that was not great. I was a faculty member and you know that was lovely. If we can call that a real job like a real job in industry.

Yeah, I know what you mean. It was, yeah, it was. And it was awesome.

Yeah. Yeah. It was, yeah, right in both good and bad, right? That there was, you know, if you looked at what we had accomplished and what we were building it was mind-blowing, right? Like, you know, and it's one of these things that there's people who work on really important technology but it's intangible, right? If you think about credit card processing, right, the rails that allow us to all commerce to occur around the world like that's incredibly interesting, really hard, deep technology. But if I tell you I work on, you know, the rails for international commerce, you're like, okay, what do you do? You know, if I tell you we're working on self-driving cars, it's like, you can understand that, right? You've got to like you have a reaction viscerally one way or another about it. You understand what driving means. You understand how important that is. And so yeah, it's just neat to be able to work on something like that.

Yeah.

Yeah, we, that was an amazing day. So Google their company all hands is called TGIF. Thank God it's Friday, even though they do it on a Thursday, which is entertaining in of itself. But it was done for good reason. It was done because, you know, they've got international, they had the fortune of being a big international company, you know, Friday, Saturday. Yeah, exactly. But it was still hilarious. And yeah, got to get up on stage in front of I don't know, probably 60,000 Googlers at the time and be like, 'Hey, we do awesome stuff at Google.' And by the way, we make cars drive themselves, right? And talk about the tech and, you know, it's a company of engineers and, you know, being able to kind of lay out some of the cool technology and, you know, it was, you know, you become an instant folk hero. Crazy. It was very cool.

I'm sure nothing leaps to mind. I can tell you one at Aurora. So when we started Aurora, we were focused on driving on like light vehicles on low speed roads, urban driving. And we knew trucking was going to be an important application, but we didn't think you could see far enough to solve it. Why? Because LiDAR, the way LiDAR works, you shoot laser beams out in the world and you basically time how long they go out, how long they come back. Speed of light is constant. So you can figure out a distance from that. The farther you shoot out, the more the beam gets dispersed, the less energy that comes back. So if you want to see further, you need to shoot more energy out. At some point, you're shooting out so much energy that you'll either, depending on the wavelength, you'll either burn people's retina, or you will boil the aqueous fluid in their eyeballs, which you don't want to do. So, not a good thing to do, right?

Kinetic energy for a truck driving at 70 mph versus a car driving in an urban environment at 15 to 25 is about 50-fold. So the amount of stopping distance is larger and all of that. And so anyway, we knew that this LiDAR technology was out there. We knew what the limit of it was and we knew that wasn't far enough to be able to drive trucks safely. And so we spent a bunch of time in the first couple of years of the company, trying to find a technology to unlock the ability to drive at high speeds on the freeway with big trucks. And we ultimately found this company in Montana called Blackmore. And they had built a special way called frequency modulated continuous wave, a special way to measure that time of flight for the light. And the way they do that is instead of looking for a bright spike to come back, they look at wave interference and they look at the phase difference in the waves and that allows you to figure out that phase difference allows you to figure out how far things are.

Totally, like it's completely completely obvious. But the magic in this and this is again in the, you know, of course I knew this, is that that self-interference thing is called self-heterodyne where you're kind of overlapping these waves and it turns out that through magic of physics and handwaving is a 10 to 20-fold amplification of the signal and so that means that with the same amount of energy you can see much further. And so when we discovered we saw this team and, you know, thought about this technology in the past we're like it's just not feasible. Met this team, Randy Reel was the founder of it. He's like, 'Oh, we're doing FMCW.' I'm like, 'Yeah, right.' And then, you know, I'll take the meeting but I don't believe you. Then got to know him, got to know the team. I was like, 'Oh, they've actually cracked this.' And so we ended up acquiring that company at Aurora in 2018. It was about 70 people, something like that. And it is one of the best acquisitions you could imagine. It's one of the fundamental bits of technology we have.

And it's one of those acquisitions, you know, you hear like all acquisitions fail, right? Or at least most acquisitions fail. And this one's the opposite of that. Like it has turned out to have been we were right about the need for the technology. We were right that these guys know how to solve it and we were right that we could industrialize it. And the teams for the most part stuck around and it's on every one of our trucks today.

And without it I don't know how, you know, like I can imagine alternatives but this one works.

Yeah. So, I left Google and I spent, I did not leave Google to start Aurora. I left Google because it was time to get out and then spent I don't know three or six months trying to figure out what to do. And so we started Aurora kind of in earnest at the beginning of 2017. And I think we acquired Blackmore.

Yeah, it was not. So our mission has been the same from day one, which is to deliver the benefits of self-driving technology safely, quickly, and broadly. But if you look at our early pitch decks, you know, when we, it was okay, we're going to do cars. And then, you know, we'll look at adjacencies like trucking and other places. And again, that was partly because we just didn't see how you could actually see far enough to drive these things.

Yeah, sounds like Waymo. Yeah.

Yeah, it certainly crossed my radar. Like I was trying to figure out what to do, right? I was pretty down and was looking at what the opportunities was and it was very flattering, right? And when I left Google it made national news which was like well that's kind of cool. And so I got a lot of interest and had a lot of really good conversations with folks. And I ended up like I was trying to think about what do I want to do next and really it came down to minimizing regret, right? I was thinking things through and I was like well I believe in this technology, it's going to happen and, you know, when I'm talking to my grandkids one day I don't want to be like, you know, I could have been a contender, I could have helped make that happen. And so, you know, when I did the math, like did the reason through it, it was, you know, so I could start a company to go do this and we could fail. And it's like, you know, right? You tried, didn't work out. That happens. Start a company, it kind of works, but I don't like it. You're like, okay, I can hand it off and leave. Or we could start a company, it's successful, and you enjoy it. And that's, you know, obviously the happy path and the one you hope for. And it seemed like all of those options were better than, you know, at some point in the future going and saying, you know, I could have done that. I could have been part of that, you know. And so that was why I ultimately decided to kind of get back into it. Also felt like I had a unique set of experiences and skills that, you know, I could have gone and been a product or programming engineering manager on something else but it would have missed chunk of value that I have, right, which is understanding the space and the technology and, you know, what's needed to make it work and so that seemed like I should go use that, right? That's like, you know, without trying to sound goofy about it but that's kind of my purpose, right, was to go do this and make it happen given the experiences I'd had.

I decided that the right path was to start a company because I'm a big believer that a well-run company executes on its mission. And that at some point things go badly and you have to put the chips down and you should put them down on the things that are aligned with your mission. And so if you're a car company, your mission is not to deliver the benefits of self-driving technology safely, quickly, and broadly. If you're Ford, it's something like, you know, enable every American to be able to buy a reasonably priced car or something like that. I don't know if that's still it, but it's that gist. You know, if you're Google, it's organize the world's information to make it universally accessible. And so at some point, this is a hard problem. Most companies are not going to have the patience that Google has. And so if we're going to go do this, need to do it in a place where we can control our destiny. And that, you know, we're not going to end up being the bet that other companies don't make. And if you look at what's happened over the last 10 years, there's been a number of these efforts, whether it's Apple shutting down Titan or Ford and Volkswagen shutting down Argo or most recently General Motors shutting down Cruise. At the end of the day, each of these companies has looked at their mission and said, 'My resources, whether they're people or dollars or time, are not being served by not serving our company's mission by doing this work.' And so that seemed like it was clear, well, it was clear to me when we started Aurora that that was going to be one of our, you know, benefits was that we knew what we were about. We were focused on this and, you know, maybe we'll fail but we at least control our destiny.

We raised 300 million, 390, 395 million in the private markets.

No, no. I feel like we've been pretty, that's pretty good. We were pretty efficient in total.

Yeah. Well, but at the time like we did our Series A was I think 90 million, 90 million raise which felt large.

That was 2017.

Yeah.

Blackmore was the company. First Light is the technology that we call it internally.

Yeah.

We SPACed in 2021.