Chris Urmson: Self-Driving Cars at Aurora, Google, CMU, and DARPA

[0] The following is a conversation with Chris Armson.

[1] He was the CTO of the Google Sub -Driving Car team, a key engineer and leader behind the Carnegie Mellon University Autonomous Vehicle Entries in the DARPA Grand Challenges and the winner of the DARPA Urban Challenge.

[2] Today, he's the CEO of Aurora Innovation, an autonomous vehicle software company.

[3] He started with Sterling Anderson, who was the former director of Tesla Autopilot and Drew Bagnell, Uber's former autonomy and perception lead.

[4] Chris is one of the top roboticist and autonomous vehicle experts in the world and a long -time voice of reason in a space that is shrouded in both mystery and hype.

[5] He both acknowledges the incredible challenges involved in solving the problem of autonomous driving and is working hard to solve it.

[6] This is the artificial intelligence podcast.

[7] If you enjoy it, subscribe on YouTube, give it five stars on iTunes, support it on Patreon, or simply connect with me on Twitter, Alex Friedman, spelled F -R -I -D -M -A -N.

[8] And now here's my conversation with Chris Armson.

[9] You were part of both the DARPA Grand Challenge and the DARPA Urban Challenge, teams at CMU with Red Whitaker.

[10] What technical or philosophical things have you learned from these races?

[11] I think the high order bit was that it could be done.

[12] I think that was the thing that was incredible.

[13] about the first of the grand challenges that I remember you know I was a grad student Carnegie Mellon and there we there was kind of this dichotomy of it seemed really hard so that would be cool and interesting but you know at the time we were the only robotics institute around and so you know if we went into it and fell on our faces that would that would be embarrassing so I think you know just having the will to go do it to try to do this thing that at the time was marked as, you know, darn near impossible.

[14] And then after a couple of tries, be able to actually make it happen.

[15] I think that was, you know, that was really exciting.

[16] But at which point did you believe it was possible?

[17] Did you, from the very beginning, did you personally?

[18] Because you're one of the lead engineer.

[19] You actually had to do a lot of the work.

[20] Yeah, I was the technical director there and did a lot of the work along with a bunch of other really good people.

[21] Did I believe it could be done?

[22] Yeah, of course, right?

[23] why would you go do something you thought was completely impossible?

[24] We thought it was going to be hard.

[25] We didn't know how we're going to be able to do it.

[26] We didn't know if we'd be able to do it the first time.

[27] Turns out we couldn't.

[28] That, yeah, I guess you have to.

[29] I think there's a certain benefit to naivete, right?

[30] That if you don't know how hard something really is, you try different things and, you know, it gives you an opportunity that others who are, you know, wiser maybe don't, don't have.

[31] What were the biggest pain points?

[32] Mechanical, sensors, hardware, software, algorithms for mapping, localization, just general perception control.

[33] But like hardware software, first of all.

[34] I think that's the joy of this field is that it's all hard and that you have to be good at each part of it.

[35] So for the first, for the urban challenges, if I look, back at it from today, it should be easy today.

[36] That, you know, it was a static world.

[37] There weren't other actors moving through it, is what that means.

[38] It was out in the desert, so you get really good GPS.

[39] You know, so that went, and, you know, we could map it roughly.

[40] And so in retrospect, now, it's, you know, it's within the realm of things we could do back then.

[41] just actually getting the vehicle and the you know there's a bunch of engineering work to get the vehicle so that we could control it and drive it that's you know that's still a pain today but it was even more so back then uh and then the uncertainty of exactly what they wanted us to do was was part of the challenge as well right you didn't actually know the track heading it you knew approximately but you know it didn't actually know the route the route that's going to be taken that's right we didn't know the route we didn't even the way the rules had been described, you had to kind of guess.

[42] So if you think back to that challenge, the idea was that the government would give us, the DARPA would give us a set of waypoints and kind of the width that you had to stay within between the line that went between each of those waypoints.

[43] And so the most devious thing they could have done is set, you know, a kilometer wide corridor across, you know, a field of scrub brush and rocks and said, you know, go figure it out.

[44] Fortunately, it really, it turned into basically driving along a set of trails, which, you know, is much more relevant to the application they were looking for.

[45] But no, it was a hell of a thing back in the day.

[46] So the legend, Red was kind of leading that effort in terms of just broadly speaking.

[47] So you're a leader now.

[48] What have you learned from Red about leadership.

[49] I think there's a couple of things.

[50] One is, you know, go and try those really hard things.

[51] That's where there is an incredible opportunity.

[52] I think the other big one, though, is to see people for who they can be, not who they are.

[53] It's one of the things that I actually, one of the deepest lessons I learned from Red was that he would look at, you know, undergraduates or graduate students and empower them to be leaders, to, you know, have responsibility, to do great things that I think another person might look at them and think, oh, well, that's just, you know, an undergraduate student.

[54] What could they know?

[55] And so I think that, you know, kind of trust but verify, have confidence in what people can become, I think is a really powerful thing.

[56] So through that, let's just like fast forward through the history.

[57] can you maybe talk through the technical evolution of autonomous vehicle systems from the first two grand challenges to the urban challenge to today?

[58] Are there major shifts in your mind or is it the same kind of technology just made more robust?

[59] I think there's been some big steps.

[60] So for the grand challenge, the real technology that unlocked that was...

[61] HD mapping.

[62] Prior to that, a lot of the off -road robotics work had been done without any real prior model of what the vehicle was going to encounter.

[63] And so that innovation, the fact that we could get decimeter resolution models was really a big deal.

[64] And that allowed us to kind of bound the complexity of the driving problem the vehicle had and allowed it to operate at speed because we could assume things about the environment that it was going to encounter.

[65] So that was a that was one of the, that was the big step there.

[66] For the urban challenge, you know, one of the big technological innovations there was the multi -beam LIDAR and be able to generate high resolution, you know, mid to long range 3D models the world and use that for you know for understanding the world around the vehicle and that was really a you know kind of a game -changing technology in parallel with that we saw a bunch of other technologies that had been kind of converging half their their day in the sun so uh baysian estimation uh had been you know slam had been a big field in robotics, you know, you would go to a conference, you know, a couple years before that in, every paper would effectively have slams somewhere in it.

[67] And so seeing that, you know, that those patient estimation techniques, you know, play out on a very visible stage.

[68] You know, I thought that was, that was pretty exciting to see.

[69] And mostly slam was done based on LIDAR at that time.

[70] Well, yeah, and in fact, we weren't really doing Slam per se.

[71] You know, it, you know, you know, in real time because we had a model ahead of time, we had a roadmap, but we were doing localization.

[72] And we're using, you know, the LiDAR or the cameras, depending on, you know, who exactly was doing it to localize to a model of the world.

[73] And I thought that was, that was a big step from kind of naively trusting GPS, INS before that.

[74] And again, like, lots of work had been going on in this field.

[75] Certainly this was not doing anything particularly innovative in Slam or in localization, but it was seeing that technology necessary in a real application on a big stage, I thought was where it's very cool.

[76] So for the Urban Challenge, those are already maps constructed offline?

[77] Yes.

[78] In general.

[79] Okay.

[80] And did people do that?

[81] The individual teams do it individually, so they had their own different approaches there, or did everybody kind of, share that information, at least intuitively.

[82] So the DARPA gave all the teams a model of the world, a, you know, a map.

[83] And then, you know, one of the things that we had to figure out back then was, and it's still one of these things that trips people up today, is actually the coordinate system.

[84] So you get a latitude, longitude, and, you know, to so many dozen places, you don't really care about kind of the ellipsoid of the earth that's being used.

[85] But when you want to get to 10 centimeter or centimeter resolution, you care whether the corner system is, you know, NADS 83 or WGS 84 or, you know, these are different ways to describe both the kind of non -sphericalness of the earth, but also kind of the act in, I think, I can't remember which one, the tectonic shifts that are happening and how to transform, you know, the global datum as a function of that.

[86] So you're getting a map and then actually matching it to reality to centimeter resolution.

[87] That was kind of interesting and fun back then.

[88] So how much work was the perception doing there?

[89] So how much were you relying on localization based on maps without using perception to register to the maps?

[90] And I guess the question is how advanced was perception at that point.

[91] It's certainly behind where we are today.

[92] We're more than a decade since the urban challenge.

[93] but the core of it was there that we were tracking vehicles.

[94] We had to do that at 100 plus meter range because we had to merge with other traffic.

[95] We were using, you know, Bayesian, again, Bayesian estimates for state of these vehicles.

[96] We had to deal with a bunch of the problems that you think of today of predicting where that vehicle is going to be a few seconds into the future.

[97] We had to deal with the fact that there were multiple hypotheses for that.

[98] because a vehicle at an intersection might be going right or it might be going straight or it might be making a left turn.

[99] And we had to deal with the challenge of the fact that our behavior was going to impact the behavior of that other operator.

[100] And we did a lot of that in relatively naive ways, but it kind of worked.

[101] Still had to have some kind of solution.

[102] And so where does that 10 years later, where does that take us today from that artificial city construction to real cities, to the urban environment.

[103] Yeah, I think the biggest thing is that the, you know, the actors are truly unpredictable.

[104] That most of the time, you know, the drivers on the road, the other road users are out there behaving well, but every once in a while they're not.

[105] The variety of other vehicles is, you know, you have all of them.

[106] In terms of behavior, terms of perception, or both?

[107] Both.

[108] that we have, you know, back then we didn't have to deal with cyclists, we didn't have to deal with pedestrians, didn't have to deal with traffic lights, you know, the scale over which that you have to operate is now, you know, as much larger than, you know, the air base that we were thinking about back then.

[109] So what, easy question, what do you think is the hardest part about driving?

[110] Easy question.

[111] Yeah.

[112] No, I'm joking.

[113] I'm sure nothing really jumps solid at you.

[114] as one thing.

[115] But in the jump from the urban challenge to the real world, is there something that's a particular you foresee a very serious difficult challenge?

[116] I think the most fundamental difference is that we're doing it for real, that in that environment, it was both a limited complexity environment because certain actors weren't there, because the roads were maintained, There were barriers keeping people separate from robots at the time.

[117] And it only had to work for 60 miles, which looking at it from, you know, 2006, it had to work for 60 miles.

[118] Looking at it from now, you know, we want things that we'll go and drive for, you know, half a million miles.

[119] And, you know, it's just a different game.

[120] So how important you said LIDAR came into the game.

[121] early on, and it's really the primary driver of autonomous vehicles today as a sensor.

[122] So how important is the role of LiDAR in the sensor suite in the near term?

[123] So I think it's essential.

[124] You know, I believe, but I also believe the cameras are essential, and I believe the radar is essential.

[125] I think that you really need to use the composition of data from these different sensors if you want the thing to really be robust.

[126] The question I want to ask, let's see if we're going to take.

[127] is what are your thoughts on the Elon Musk provocative statement that LIDAR is a crutch, that is a kind of, I guess, growing pains, and that much of the perception task can be done with cameras.

[128] So I think it is undeniable that people walk around without, you know, lasers in their foreheads.

[129] And they can get into vehicles and drive them.

[130] And so there's an existence proof.

[131] that you can drive using, you know, passive vision.

[132] No doubt, can't argue with that.

[133] In terms of sensors, yeah.

[134] Yeah, in terms of sensors, right?

[135] So, like, there's an example that, you know, we all go do it many of us every day.

[136] In terms of LIDAR being a crutch, sure.

[137] But, you know, in the same way that, you know, the combustion engine was a crutch on the path to an electric vehicle.

[138] in the same way that, you know, any technology ultimately gets replaced by some superior technology in the future.

[139] And really what the way that I look at this is that the way we get around on the ground, the way that we use transportation is broken.

[140] And that we have, you know, this, you know, what was, I think the number I saw this morning, 37 ,000 Americans killed last year on our roads.

[141] And that's just not acceptable.

[142] And so any technology that we can bring to bear that accelerates this, you know, self -driving technology coming to market and saving lives is technology we should be using.

[143] And it feels just arbitrary to say, well, you know, I'm not okay with using lasers because that's whatever, but I am okay with using an 8 megapixel camera or a 16 megapixel camera.

[144] You know, like it's just these are just bits of technology and we should be taking.

[145] the best technology from the tool bin that allows us to go and solve a problem.

[146] The question I often talk to, well obviously you do as well to the sort of automotive companies and you know if there's one word that comes up more often than anything is cost and trying to drive cost down.

[147] So while it's true that it's a tragic number of the 37 ,000, the question is what and I'm not the one asking this question because I hate this question, but we want to find the cheapest sensor suite that creates a safe vehicle.

[148] So in that uncomfortable trade -off, do you foresee LIDAR coming down in cost in the future, or do you see a day where level four autonomy is possible without LIDAR?

[149] I see both of those, but it's really a matter of time.

[150] And I think really, maybe I would talk to the question you asked about, you know, the cheapest sensor.

[151] I don't think that's actually what you want.

[152] What you want is a sensor suite that is economically viable.

[153] And then after that, everything is about margin and driving cost out of the system.

[154] What you also want is a sense suite that works.

[155] And so it's great to tell a story about how, you know, how it would be better to have a self -trivate system with a $50 sensor instead of a $500 sensor.

[156] But if the $500 sensor makes it work and the $50 sensor doesn't work, you know, who cares?

[157] So long as you can actually, you know, have an economic opportunity, you know, there's an economic opportunity there.

[158] And the economic opportunity is important because that's how you actually have a sustainable business.

[159] And that's how you can actually see this come to scale and be out in the world.

[160] And so, when I look at LiDR, I see a technology that has no underlying fundamentally, you know, expense to it, fundamental expense to it.

[161] It's going to be more expensive than an imager because, you know, CMOS processes or, you know, fab processes are dramatically more scalable than mechanical processes.

[162] But we still should be able to drive cost down substantially on that side.

[163] And then I also do think that with the right business model, you can absorb more, you know, certainly more cost on the bill on materials.

[164] Yeah, if the census suite works, extra value is provided, thereby you don't need to drive cost down to zero.

[165] It's basic economics.

[166] You've talked about your intuition that level two autonomy is problematic because of the human factor of vigilance, that command, complacency, overtrust, and so on, just thus being human.

[167] We overtrust the system.

[168] We start doing, even more so partaking in the secondary activities like smartphone and so on.

[169] Have your views evolved on this point in either direction?

[170] Can you speak to it?

[171] So, and I want to be really careful because sometimes this gets twisted in a way that I certainly didn't intend.

[172] So active safety systems are a really important technology that we should be pursuing and integrating into vehicles.

[173] And there's an opportunity in the near term to reduce accidents, reduce fatalities, and that's, and we should be, we should be pushing on that.

[174] Level two systems are systems where the vehicle is controlling two axes.

[175] So, you know, breaking and, breaking and thrall slash steering.

[176] And I think there are variants of level two systems that are supporting the driver that absolutely, like, we should, we should encourage to be out there.

[177] Where I think there's a real challenge is in the human factors part around this and the misconception from the public around the capability set that that enables and the trust that they should have at it.

[178] And that is where I kind of, I am actually incrementally more concerned around level three systems and how exactly a level two system is marketed and delivered.

[179] and how much effort people have put into those human factors.

[180] So I still believe several things around this.

[181] One is people will overtrust the technology.

[182] We've seen over the last few weeks, you know, a spate of people sleeping in their Tesla.

[183] You know, I watched an episode last night of Trevor Noah talking about this.

[184] And, you know, him, you know, this is a smart guy who has a lot of resources at his disposal describing a Tesla as a self -driving car and that why shouldn't people be sleeping in their Tesla?

[185] It's like, well, because it's not a self -driving car and it is not intended to be and, you know, these people will almost certainly, you know, die at some point or hurt other people.

[186] And so we need to really be thoughtful about how that technology is described and brought to market.

[187] I also think that because of the economic issue, you know, economic challenges we were just talking about, that that technology path will, these level two driver assistance systems, that technology path will diverge from the technology path that we need to be on to actually deliver truly self -driving vehicles, ones where you can get in and sleep and have the equivalent or better safety than, you know, a human driver behind the wheel.

[188] because again the economics are very different in those two worlds and so that leads to divergent technology so you just don't see the economics of gradually increasing from level two and doing so quickly enough to where it doesn't cost safety critical safety concerns you believe that it needs to diverge at this point into different basically different routes And really that comes back to what are those L2 and L1 systems doing?

[189] And they are driver assistance functions where the people that are marketing that responsibly are being very clear and putting human factors in place such that the driver is actually responsible for the vehicle and that the technology is there to support the driver.

[190] And the safety cases that are built around those are dependent.

[191] on that driver attention and attentiveness.

[192] And at that point, you can kind of give up to some degree, for economic reasons, you can give up on, say, false negatives.

[193] And the way to think about this is for a foreclosure mitigation braking system, if it, half the times the driver missed a vehicle in front of it, it hit the brakes and brought the vehicle to a stop, that would be an incredible, incredible advance in safety on our roads, right?

[194] That would be equivalent to seatbelts.

[195] But it would mean that if that vehicle wasn't being monitored, it would hit one out of two cars.

[196] And so economically, that's a perfectly good solution for a driver assistance system.

[197] What you should do at that point, if you can get it to work 50 % of the time, is drive the cost out of that so you can get it on as many vehicles as possible.

[198] But driving the cost out of it doesn't drive up performance on.

[199] the false negative case.

[200] And so you'll continue to not have a technology that could, you know, really be available for a self -driven vehicle.

[201] So clearly the communication, and this probably applies to all four vehicles as well, the marketing and a communication of what the technology is actually capable of, how hard it is, how easy it is, all that kind of stuff, is highly problematic.

[202] So say everybody in the world was perfectly communicated and were made to be completely aware of every single technology out there what they what it's able to do what's your intuition and now we're maybe getting into philosophical ground is it possible to have a level two vehicle where we don't over trust it I don't think so if people truly understood the risks and internalized it then then sure you could do that safely but that that's a world that doesn't exists that people are going to they're going to you know if the facts are put in front of them they're going to then combine that with their experience and you know let's say they're they're using an l2 system and they go up and down the 101 every day and they do that for a month and it just worked every day for a month like that's pretty compelling at that point you know just even if you know the statistics you're like well I don't know maybe there's something little funny about those.

[203] Maybe they're, you know, driving in difficult places.

[204] Like, I've seen it with my own eyes.

[205] It works.

[206] And the problem is that that sample size that they have, so it's 30 miles up and now.

[207] So 60 miles times 30 days, 60, 180, 1 ,800 miles.

[208] That's a drop in the bucket compared to the one, you know, what, 85 million miles between fatalities.

[209] And so they don't really have a true estimate based on their personal experience of the real risks, but they're going to trust it anyway because it's hard not to It worked for a month.

[210] What's going to change?

[211] So even if you start a perfect understanding of the system, your own experience will make it drift.

[212] I mean, that's a big concern.

[213] Yeah.

[214] Over a year, over two years, even.

[215] It doesn't have to be months.

[216] And I think that as this technology moves from what I would say is kind of the more technology savvy ownership group to, you know, the mass market, you may be able to have some of those folks who are really familiar with technology.

[217] be able to internalize it better, and your kind of immunization against this kind of false risk assessment might last longer, but as folks who aren't as savvy about that, you know, read the material and they compare that to their personal experience, I think there that, you know, it's going to, it's going to move more quickly.

[218] So your work, the program that you've created at Google and now at Aurora is focused more on the second path of creating full autonomy.

[219] So it's such a fascinating, I think it's one of the most interesting AI problems of the century, right?

[220] I just talk to a lot of people, just regular people, I don't know, my mom about autonomous vehicles.

[221] And you begin to grapple with ideas of giving your life control over to a machine.

[222] It's philosophically interesting.

[223] It's practically interesting.

[224] So let's talk about safety.

[225] How do you think?

[226] We demonstrate, you've spoken about metrics in the past, how do you think we demonstrate to the world that an autonomous vehicle, an Aurora system is safe?

[227] This is one where it's difficult because there isn't a soundbite answer.

[228] That we have to show a combination of work that was done diligently and thoughtfully, and this is where something like a functional safety process is part of that is like, here's the way we did the work, That means that we were very thorough.

[229] So, you know, if you believe that we, what we said about this is the way we did it, then you can have some confidence that we were thorough in the engineering work we put into the system.

[230] And then on top of that, you know, to kind of demonstrate that we weren't just thorough, we were actually good at what we did, there will be a kind of a collection of evidence in terms of demonstrating that the capabilities work the way we thought they did, you know, statistically and to whatever, degree we can demonstrate that both in some combination of simulation, some combination of unit testing and decomposition testing, and then some part of it will be on -road data.

[231] And I think the way we'll ultimately convey this to the public is there'll be clearly some conversation with the public about it, but we'll kind of invoke the kind of the trusted nodes and that we'll spend more time being able to go into more depth with folks like NHTSA and other federal and state regulatory bodies and kind of given that they are operating in the public interest and they're trusted that if we can show enough work to them that they're convinced then you know I think we're in a in a pretty good place that means you work with people that are essentially experts at safety to try to discuss and show do you think the answer is probably no but just in case Do you think there exists a metric?

[232] So currently people have been using number of disengagement.

[233] And it quickly turns into a marketing scheme to sort of you alter the experiments you run to.

[234] I think you've spoken that you don't like.

[235] Don't love it.

[236] No, in fact, I was on the record telling DMV that I thought this was not a great metric.

[237] Do you think it's possible to create a metric, a number that could demonstrate safety outside of fatalities?

[238] So I do.

[239] And I think that it won't be just one number.

[240] So as we are internally grappling with us, and at some point we'll be able to talk more publicly about it, is how do we think about human performance in different tasks, say detecting traffic lights or safely making a left turn across traffic?

[241] And what do we think the failure rates are for those different capabilities?

[242] for people and then demonstrating to ourselves and then ultimately folks in regulatory role and then ultimately the public that we have confidence that our system will work better than that.

[243] And so these individual metrics will kind of tell a compelling story ultimately.

[244] I do think at the end of the day, what we care about in terms of safety is life saved and injuries reduced.

[245] and then ultimately, you know, kind of casualty dollars that people aren't having to pay to get their car fixed.

[246] And I do think that you can, you know, in aviation they look at a kind of an event pyramid where, you know, a crash is at the top of that and that's the worst event, obviously.

[247] And then there's injuries and, you know, near miss events and whatnot and, you know, violation of operating procedures.

[248] and you kind of build a statistical model of the relevance of the low severity things, or the high severity things.

[249] I think that's something we'll be able to look at as well.

[250] Because, you know, an event per 85 million miles is a statistically a difficult thing, even at the scale of the U .S. to kind of compare directly.

[251] And that event, a fatality that's connected to an autonomous vehicle is significantly, at least currently magnified in the amount of attention to get.

[252] So that speaks to public perception.

[253] I think the most popular topic about autonomous vehicles in the public is the trolley problem formulation, right?

[254] Which has, let's not get into that too much, but is misguided in many ways.

[255] But it speaks to the fact that people are grappling with this idea of giving control over to a machine.

[256] So how do you win the hearts and minds of the people that autonomy is something that could be a part of their lives?

[257] I think you let them experience it, right?

[258] I think it's right.

[259] I think people should be skeptical.

[260] I think people should ask questions.

[261] I think they should doubt because this is something new and different.

[262] They haven't touched it yet.

[263] And I think it's perfectly reasonable.

[264] And, but at the same time, it's clear there's an opportunity to make the road safer.

[265] It's clear that we can improve access to mobility.

[266] It's clear that we can reduce the cost of mobility.

[267] And that once people try that and are, you know, understand that it's safe and are able to use in their daily lives, I think it's one of these things that will just be obvious.

[268] And I've seen this practically in, you know, demonstrations that I've, you know, given.

[269] where I've had people come in and, you know, they're very skeptical and they, again, the vehicle, you know, my favorite one is taking somebody out on the freeway and we're on the 101 driving at 65 miles an hour.

[270] And after 10 minutes, they kind of turn and ask, is that all it does?

[271] And you're like, it's a self -driving car.

[272] I'm not sure exactly what you thought it would do, right?

[273] But they, you know, it becomes mundane, which is exactly what you want a technology like this to be.

[274] Right.

[275] We don't really.

[276] When I turn the light switch on in here, I don't think about the complexity of, you know, those electrons, you know, being pushed down a wire from wherever it was and being generated.

[277] It's like I just get annoyed if it doesn't work, right?

[278] And what I value is the fact that I can do other things in this space.

[279] I can, you know, see my colleagues.

[280] I can read stuff on a paper.

[281] I can, you know, not be afraid of the dark.

[282] And I think that's what we want this technology to be like.

[283] It's in the background and people get to have those life experiences and do so safely.

[284] So putting this technology in the hands of people speaks to scale of deployment.

[285] So what do you think the dreaded question about the future because nobody can predict the future?

[286] But just maybe speak poetically about when do you think we'll see a large scale deployment of autonomous vehicles, thousand those kinds of numbers we'll see that within ten years I'm pretty confident we what's an impressive scale what moment so you've done DARPA challenge where there's one vehicle at which moment does it become wow this is serious scale so I think the moment it gets serious is when we really do have a driverless vehicle operating on roads and that we can do that kind of continuously without a safety driver without a safety driver in the vehicle i think at that moment we've we've kind of crossed the zero to one threshold and then it is about how do we continue to scale that how do we build the right business models how do we build the right customer experience around it so that it is actually you know a useful product out in the world.

[287] And I think that is really, at that point it moves from a, you know, what is this kind of mixed science engineering project into engineering and commercialization and really starting to deliver on the value that we all see here and, you know, and actually making that real in the world.

[288] What do you think that deployment looks like?

[289] Where do we first see the inkling of no safety driver, one or two cars here and there?

[290] Is it on the highway?

[291] is it in specific groups in the urban environment?

[292] I think it's going to be urban, suburban type environments.

[293] You know, with Aurora, when we thought about how to tackle this, it was kind of invoked to think about trucking as opposed to urban driving.

[294] And, you know, the, again, the human intuition around this is that freeways are easier to drive on.

[295] Because everybody's kind of going in the same direction.

[296] you know, the lanes are a little wider, et cetera.

[297] And I think that that intuition is pretty good, except we don't really care about most of the time.

[298] We care about all of the time.

[299] And when you're driving on a freeway with a truck, say, 70 miles an hour, and you've got 70 ,000 pound load with you, that's just an incredible amount of kinetic energy.

[300] And so when that goes wrong, it goes really wrong.

[301] And that those challenges that you see occur more rarely, so you don't get to learn as quickly, and they're, you know, incrementally more difficult than urban driving, but they're not easier than urban driving.

[302] And so I think this happens in moderate speed urban environments because there, you know, if two vehicles crash at 25 miles per hour, it's not good, but probably everybody walks away.

[303] And those events where there's the possibility for that occurring happen frequently so we get to learn more rapidly we get to do that with lower risk for everyone and then we can deliver value to people that they need to get from one place to another and then once we've got that solved then the kind of the freeway driving part of this just falls out but we're able to learn more safely more quickly in the urban environment so 10 years and then scale 20 30 year I mean who knows if if a sufficiently compelling experience is created it could be faster and slower or do you think there could be breakthroughs and what kind of breakthroughs might there be that completely change that timeline?

[304] Again, not only am I asking to predict the future, I'm asking you to predict breakthroughs that haven't happened yet.

[305] So what's the, I think another way to ask that would be if I could wave a magic wand, what part of the system would I make work today to accelerate it as quick as possible, quickly as possible?

[306] Don't say infrastructure.

[307] Please don't say infrastructure.

[308] No, it's definitely not infrastructure.

[309] It's really that perception forecasting capability.

[310] So if tomorrow you could give me a perfect model of what is happening and what will happen and what will happen for the next five seconds around a vehicle on the roadway, that would accelerate things pretty dramatically.

[311] Are you, in terms of staying up at night, are you mostly bothered by cars, pedestrians, or cyclists?

[312] So I worry most about the vulnerable road users.

[313] about the combination of cyclists and cars, right?

[314] Just cyclists and pedestrians because, you know, they're not in armor.

[315] You know, the cars, they're bigger.

[316] They've got protection for the people.

[317] And so the ultimate risk is lower there.

[318] Whereas a pedestrian or a cyclist, they're out in the road and, you know, they don't have any protection.

[319] And so, you know, we need to pay extra attention to that.

[320] Do you think about a very difficult technical challenge?

[321] of the fact that pedestrians if you try to protect pedestrians by being careful and slow they'll take advantage of that so the game theoretic dance yeah does that worry you of how from a technical perspective of how we solve that because as humans the way we solve that it's kind of nudge our way through the pedestrians which doesn't feel from a technical perspective as a appropriate algorithm but do you think think about how we solve that problem?

[322] Yeah, I think there's, I think that was actually, there's two different concepts there.

[323] So one is, am I worried that because these vehicles are self -driving, people kind of step in the road and take advantage of them?

[324] And I've heard this and I don't really believe it because if I'm driving down the road and somebody steps in front of me, I'm going to stop.

[325] Right?

[326] Like, even if I'm annoyed, I'm not going to just drive through a problem.

[327] person stood in the road.

[328] And so I think today people can take advantage of this, and you do see some people do it.

[329] I guess there's an incremental risk because maybe they have lower confidence that I'm going to see them than they might have for an automated vehicle, and so maybe that shifts it a little bit.

[330] But I think people don't want to get hit by cars.

[331] And so I think that I'm not that worried about people walking out of the 101 and, you know, creating chaos more than they would today.

[332] regarding kind of the nudging through a big stream of pedestrians leaving a concert or something I think that is further down the technology pipeline I think that you're right that's tricky I don't think it's necessarily I think the algorithm people use for this is pretty simple right it's kind of just move forward slowly and if somebody's really close then stop and and I think that that probably can be replicated pretty easily and particularly given that it's you don't do this at 30 miles an hour you do it at one that even in those situations the risk is relatively minimal but i you know it's not something we're we're thinking about in any serious way and probably the that's less an algorithm problem and more creating a human experience so the hcii people that create a visual display that you're pleasantly as a pedestrian nudged out of the way yes that's a that's a yeah that's an experience problem not an algorithm problem who's the main competitor to Aurora today?

[333] And how do you out -compete them in the long run?

[334] So we really focus a lot of what we're doing here.

[335] I think that, you know, I've said this a few times that this is a huge, difficult problem, and it's great that a bunch of companies are tackling it, because I think it's so important for society that somebody gets there.

[336] So we, you know, we don't spend a whole lot of time thinking tactically about who's out there and how do we beat that person individually.

[337] What are we trying to do to go faster ultimately?

[338] Well, part of it is the leadership team we have has got pretty tremendous experience.

[339] And so we kind of understand the landscape and understand where the cul -de -sacs are to some degree and we try and avoid those.

[340] I think there's a part of it, just this great team we've built.

[341] People, this is a technology and a company that people believe in the mission, of, and so it allows us to attract just awesome people to go work.

[342] We've got a culture, I think, that people appreciate that allows them to focus, allows them really spend time solving problems, and I think that keeps them energized.

[343] And then we've invested heavily in the infrastructure and architectures that we think will ultimately accelerate us.

[344] So because of the folks we're able to bring in early on, because of the, the, you know, the great investors we have, you know, we don't spend all of our time doing demos and kind of leaping from one demo to the next.

[345] We've been given the freedom to invest in infrastructure to do machine learning, infrastructure to pull data from our on -road testing, infrastructure to use that to accelerate engineering.

[346] And I think that early investment and continuing investment in those kind of tools will ultimately allow us to accelerate and do something pretty incredible.

[347] Chris, beautifully put.

[348] It's a good place to end.

[349] Thank you so much for talking today.

[350] Thank you very much.

[351] Really enjoyed it.