Episode 7 | Jun 30, 2025 - Automation Accelerates

Billy Riggs (00:41)
I'm Billy Riggs, if you've not tuned in before, and I'm joined by my co-host, Vipul Vyas. Today we have the amazing co-director of the Autonomous Vehicles and the City Initiative, Dr. Henriette Cornet, joining us to talk autonomous vehicle systems.

Vipul you want to add anything else about what you hope to get out of this, but I definitely hope to hear about Dr. Cornet's recent trip to Hamburg at a global conference. The UITP conference where they were talking about global trends in this industry.

Vipul Vyas (01:14)
I think you guys are the experts, so I'm hoping to learn on this call, even though I can't help myself and we'll talk more than I should.

Billy Riggs (01:22)
She's always good for a knowledge bomb. Dr. Cornet?

Henriette Cornet (01:26)
Yeah, and thanks for having me. very looking forward to the discussion.

Billy Riggs (01:29)
Okay, well, maybe you could just start reminding us who you are and a little bit about your background in not only the the AV industry, but a little bit about your your background, working with with OEMs. And particularly, I know you've had a broad spectrum of experience. But you did work in in in a little bit in the in sustainability space as well.

and battery tech and stuff like that. And we are interested in jobs across and the economy across the AV, EV spectrum. And maybe you could just talk about your experience working in the vehicle tech space.

Henriette Cornet (02:04)
Yeah, yeah, yeah, sure. I'm happy to. So yeah, Henriette Cornet, so same with a French accent. And so I'm trained as an engineer and quickly after my master, I got very interested in the topic of sustainability actually. So I moved to Germany and there I did my PhD, this, focusing on innovation for public transport that can help having positive impact on like...

environmental aspects or societal issues. So really sustainability as a whole, thinking of economic viability for city environment and the society itself as a people. So that's why I got really much into the transportation area at this moment, still having in mind the sustainability aspect. And step by step along my career, I was interested as well in innovations. I believe that innovations are key and

always looking forward, always thinking of what's ahead and what can we improve in our systems, not making tech for the sake of technology, but really seeing how can it improve our lives. And I work for the automotive industry, as you mentioned, in Germany, in the field of electric, like electric mobility. And I learn a lot there. I could see also the mindset how car manufacturers think and what...

what are the goals, etcetera, and how can they comply as well? So we may think that they only want to sell cars, which are their main business, of course, but they have more and more pressure also to comply with environmental regulation and also from customers, they want more safety, they want less environmental impact. So that's also something I learned there. But I got kind of bored by electrification at some point, and I saw the field of autonomous mobility coming.

And I really thought, there's something there. There's something that can change the entire society and the way we move, especially in cities with this technology of autonomous driving. And that's where I moved to Singapore, leading a research group on design for autonomous mobility. So really the design aspect, how human will interact with AVs in the future, inside the vehicle and outside the vehicle. Super interesting, focusing on their needs, expectations and so on. And then moving back to Europe, I led a very large

European funded project that deployed vehicles like AVs, mainly shuttles. if you're, I don't know if your audience is familiar with shuttles, with autonomous shuttles, it would be like the EasyMile and Navia from back then, let's where you have like 10, 12 passenger capacity and they are very low speed and they have like some specific, like restricted use case, let's say. But

quite helpful for cities to get familiar with the technology. And that's what we did with this huge project. I say huge because there were 15 European countries involved in the project, something like 450 researchers, I mean, people working on that and 70 vehicles deployed all over Europe. So it was a lot of insights gained. And that's where I attended for the first time this summit, the UITP summit that you mentioned. And my role there was really to push

public transit agencies to look at this technology and make sure that the discussion is not only led by automotive industry, that really public transport players know what's happening. Maybe if they don't develop the technologies themselves, at least they can see the usage of that technology to serve their customers. that's since then and after I moved to US, but since then I kept exactly this mindset in my work.

on advising cities, advising public transit agency on how to make the best use of technology. And that's what I'm doing now in the US. I have my own company called Urban Innovate, focusing exactly on that as well as on travel behaviors and how people choose different modes. And I'm also an adjunct professor at the University of San Francisco and co-directing with you, Billy, the AV & the City initiative. It was a very long intro, sorry.

Billy Riggs (06:09)
That's

fantastic. And we got lucky that we were able to slice away some of your time from directing or developing these cool projects or products, you know, along with cool innovators like Vipul. So, yeah, thanks for spending, a little morsel of your of your talent with us at the Autonomous Vehicles in the City initiative.

Vipul, any questions for Henriette before we dive into her recent experience representing us at the UITP Summit in Hamburg?

Vipul Vyas (06:38)
Well, know, we talked about the role, Henriette mentioned the role of the actual automobile manufacturers. But what I would jump to mind for me was really it's one of the few industries, I mean, it's actually not true. have the airline industry as well where the industry is really dependent on public infrastructure and public infrastructure is what makes it even possible. Otherwise, industries would not be.

viable and I don't know if that's often recognized like cars are a part of our lives because we've decided to invest massively to make them integral part of our lives and that decision can be changed and I don't think people totally grasp that I don't think maybe even the industry doesn't totally grasp that because it's been a hundred plus years since they themselves

Henriette Cornet (07:19)
Mm-hmm.

Vipul Vyas (07:30)
understood that with regard to other industries like light rail and helped and you know other radical forms of mobility such as walking and they themselves kind of uprooted all that. So I think that is an interesting place to sort of for me to learn more is how that has changed and what change is possible and what change may just not be possible from a practical perspective as well.

Henriette Cornet (07:56)
Yeah, that's a big one to start, I would say, because it has this component of public in terms of governmental decision and somehow politics as well behind and then the industry and all the economic benefits that they want to get. And there are some tensions, let's say, that exist. And if you take the people with the needs, people are also taxpayers, people are also voters.

So it makes this entire equilibrium kind of, it's not always, I think, like decisions on infrastructure are not always made on long-term thinking, unfortunately. And it's more like to please the very like next needs that comes and building more roads, more roads, more roads. I think there will be some limit at some point. And US being so broad and having so much space.

doesn't see that, in Europe we can see scarcity of space and that's where probably earlier we needed to think of other ways to move people around because we saw the need of building more more roads.

Billy Riggs (08:59)
Yeah.

Vipul Vyas (09:00)
I wonder if it's a scarcity.

I actually think it's also a poverty of imagination. And you see that in...

people's reflexive reactions, whether that's in San Francisco at the closing of the Great Highway, whether that's congestion pricing in New York, there's an assumption that cars are so integral to our lives that they're essentially indispensable and the idea of dispensing with them is morally outrageous and repugnant and

Henriette Cornet (09:29)
Mm-hmm.

Vipul Vyas (09:31)
That's because we have so solidified our physical built environment, our culture, everything around this one device that it is tough to get anyone to think any other way. I mean, we have...

the high speed rail project in California, is probably poorly managed, has cost overruns. But most transit projects do, to be honest. But they don't get the kind of level of scrutiny because rail is viewed as a subsidy and roads are viewed as an investment. really, mean, the Dutch, think people, I'll just say this for the benefit of people who are less aware.

Henriette Cornet (10:02)
Mm-hmm.

Vipul Vyas (10:12)
the western northern europeans in many cases northern germany all the way into the netherlands people well those cities were old that's why they were like the way they are and blah blah blah you know most of them were completely destroyed during world war two so they're not necessarily that old per se and they actually had a very us-like development pattern until they decided they didn't like it

Henriette Cornet (10:29)
Mm. ⁓

Vipul Vyas (10:38)
and the Dutch I think made a conscious choice. So the only thing I'll just say is my final thought there is, there's conscious choices that are required. And if you don't understand what's in your possibility set, you're not gonna feel like you have as many choices as you do. And that's what I think the US is characterized by, generally.

Henriette Cornet (10:56)
Mm-hmm.

Vipul Vyas (10:59)
a decision.

Billy Riggs (11:01)
Yeah. So Henriette, do you want to talk to us about what's happening in Hamburg and maybe what you learned, what you observed? Is there anything exciting coming down the pike in Europe that we can learn about? And What does Europe need to learn from us?

Henriette Cornet (11:05)
Thank

Yeah, that's kind of exactly what I've been focusing on these past years, having collaborated with US when I was in Europe and now vice versa. So yeah, it's very different approach with the two continents, let's say even if you take North America as a largest part. And in US we had this huge, massive private investment coming from company like Google, Amazon, et cetera.

which have enabled over the years, we did not happen overnight, but over the years to build company as robust as Waymo with an exceptional driving capability and this Waymo driver being really leading kind of what's happening now in terms of technology and capabilities on the road and safety mainly. think that's the point I hope we'll talk again about this aspect of safety. And we don't have, we have never had this type of investment in Europe. We are struggling finding

putting together a lot of money, a lot of investments. Mostly everything that has happened in Europe was funded by the member states themselves, Germany and France, I would say, would be leaders. They have put a lot of public money into research and development, as well as the European Commission. But still, if you remember, I mentioned this project, this large project, we are talking about 30 million euros.

that are spread over 70 partners, entities. So at the end, there is nothing much left to develop anything. So that's the biggest kind of historical, we could go even deeper into history, but like the short-term history and why we don't see, we don't have a Waymo players in Europe is really a matter of budget and investments that was not there. So what we do in Europe,

basically because still the players there can see some of the players at least see opportunity in this technology. The approach is rather on thinking about services, which I also find actually more interesting because it's more about, okay, the technology will come at some point where it can come from US, it can come from China, but what we will do with that? What we will do and which type of, which vehicles will it be?

put on like placed on, do we need again shuttles? Should we think again this topic of shuttles? Should it be robot taxi serving as feeder as a first last mile option? Or should it be buses? Autonomous buses was also a topic we investigated because, and it's very much public transit kind of mindset thinking of there's two components when you think of autonomous buses. One is the driver shortage that is

is a situation in US but also in Europe. With the demographic and with many different reasons, we have more and more difficulty, public transport agencies and operators have more and more difficulty finding Buzz drivers. And the other one is the cost of operation that are linked with the wage of the drivers, so to say. And the example we often give, if a city has an operator deploying in a city has like a

a fleet, large fleet of buses, approximately 40 to 50 % of the total cost of operations are linked with the wage of the drivers. And when I say that, I'm not saying remove the driver and you get the 40, 50 % kind of cash or like money you can use somewhere else because of course you will need people to do other job linked with autonomous mobility, but not having this 40 to 50 % of your total cost

linked with the wages of drivers gives you flexibility in spending the money at other places and other jobs, definitely, but in a more flexible way. And we hope also that a part of it can be reinvested in having maybe more vehicles, more frequency, more reliable services, etcetera. So you can make a transfer of cost that is not consumed entirely, like so to say this part.

by the drivers. I hope I made myself clear because I know it's a very very touchy topic and I don't mind, the mindset is really not of saying let's get rid of the driver because it costs too much, it's not that, it's on rethinking how to allocate the cost in a better way.

So back to your question Billy on what's happening in Hamburg, now that I've said it a bit, these two different mindset and like historical backgrounds from US and Europe. In Hamburg,

And since it was a conference organized by the International Association of Public Transport, of course, it was very much focused on this very topic, which is the topic I found most interesting in this area these days. And I was happy in comparison with the edition of two years ago to see more of these innovations, more players going towards autonomous vehicles. And the two that are worth mentioning would be MOIA.

MOIA has been all over the place in this summit; so MOIA is based in Hamburg and what they are doing basically is they are operating an on-demand service in Hamburg and it's a shared mobility in terms of pooling. So they are very very focused on pooling people and they want to prove that it's possible and that people will accept to share a vehicle with strangers and the way they do that so first they started manually

And the way they do that, have like van, mini van, where you have a space inside that you are not shoulder and shoulder with someone you don't know, like Uber maybe tried some time ago with the pooling. So the vehicles are more spacious and it seems that there is an acceptance towards sharing the assets itself during the ride. And MOIA is gradually moving towards more autonomous capabilities and they collaborate with

Volkswagen, they're using the ID Buzz, so the autonomous ID Buzz, and replacing gradually, or deciding on which area it makes more sense to have these autonomous vehicles on the road, but MOIA is really doing the dispatching, the booking with the customer, etcetera, so there is kind of an intelligence gathering, the intelligence on their area, that they displayed a lot of...

progress on stage so that was great to see them and mostly see their vision I think it's inspiring for cities. Something unique about them is also that they are fully integrated to public transport meaning it's considered as part of the public transport. They are in close collaboration with Hamburger Hochbahn the agency there so it's very nice to see them working together. I haven't seen something like that in US ever so like this close collaboration.

between governmental agency and private players. And the other one I want to mention that is relevant, I think, for US as well is a company called Holon. they are, so to say for me, will be like the shuttles are back. It seems, it seems, let's see. It's still a startup and they are still doing their best and they will face a lot of issues, I'm sure. But as they want to put back like shuttles on the road and...

I'm hopeful for them because I think the form of having this small 10, 12 people shuttles is meaningful for cities. There can be a lot of use case where this makes sense. And if they have a better driving technology, which they do now, they are collaborating with Mobileye, which has proven to be solid, not yet maybe the level of Waymo, but they have driven many, many miles already.

It could be promising for the future. They operate already. They have some pilots in Germany. And the big thing is that they are building a factory in Jacksonville, Florida. So really to comply with all the Buy America and everything, all the things happening there. And Jacksonville, the City of Jacksonville, Jacksonville Transportation Authority, JTA, has already committed to buy 15 shuttles, I believe.

and the first 100 one, or they have made some commitments with them to buy them as soon as they come out of the factory, which would happen in 2027. Always a bit of waiting time. But I'm really looking forward to see these vehicles on the road and coming, yeah, kind of coming back, coming back of shuttles.

Billy Riggs (19:28)
Well, I think That's cool, Henriette. We'll see. Shuttles have their place, right? And think we talk a lot, we talk a lot in the Autonomous Vehicles in the City initiative, about how there's room for all of these tools in a public transportation system, whether or not it's a,

Henriette Cornet (19:31)
shuttle fan, I

Billy Riggs (19:46)
a for a high capacity train or or a bus, a legacy bus system. But there's also room for autonomous shuttles and, shared robotaxi services like we see in what Waymo and and formerly Cruise were providing. And I think what's what we haven't seen in Europe are these lightweight shared robotaxi solutions. And I would say, where's that in Europe? And

And I think my question is that, why, how do we, how do we get to this, this lightweight in, in Europe, we've seen, AVs treated as public infrastructure. And in US we've seen it as VC funded private services. And that pathway has accelerated AV technology much quicker in the United States.

Henriette Cornet (20:26)
Thank you.

Billy Riggs (20:36)
I should be cautious because we know Waymo's been at this game for a long time. Waymo's been at this game for 15 years. 15 years, 2012. a long time. so long story short, is that durable capital, that huge chunk of $11 to $15 billion is sitting there compared to

30 million euro that went to your SHOW project and 60 million euro that's going to this ULTIMO project. that's such a small chunk of change compared to for developing these lightweight shuttle projects. And so how do you get from zero to one with that chunk of capital for shuttle projects when you have these robotaxi projects that are able to accelerate so quickly?

Henriette Cornet (21:03)
Thank you.

Yeah

I mean, two things here. I remember Waymo mentioning that ride hailing services was the quickest way to get as many miles as possible on the road. So it was very strategically. I don't think my vision from Waymo and I'm not working for them. don't know, but following what they have been doing, I don't think they pursue any services in particular. They really want to have the best driver on the planet, so to say. I don't want to say they don't care, but like if it is

for public services, if it is for private use, that may not be the final interest per se. They don't have this type of vision, I would say, but they want to have the best driver, which they have so far. Let's see how it goes.

And the other aspect, because you mentioned the large investments and the investment difference between Europe and US, what will come, it will be kind of a dilemma in a way, and the discussions in Hamburg was also around there. We're also one there. Ever, we accept that we will never have a driver, like an autonomous driver technology in Europe. And we commit to...

by Chinese or US technology. I think some years ago, we were really much relying on US technology and very happy to do so and thinking, with China, let's see how things are going, et cetera. With everything that has happened since, let's say, last November and the changes globally, geopolitically, I can see in Europe,

tendency to say, okay, we can't rely on us anymore for tech. And this may have actually a positive effect of Europe waking up and thinking, okay, maybe we do need our own tech. And maybe we really need to do something there. And there is a mechanism that exists that can put all the member states together. It's called an IPCEI, I believe we will need to put in the reference of this episode exactly as an acronym, what is 10 for it's a

It's a project, it's a funding mechanism coordinated again by the European Commission for very, very large scale project. And there we are talking of billions of euros, of several billions of euros. A good example and a success stories behind this type of funding mechanisms is Airbus. Airbus started like that, where they say, we don't want to buy Boeing or we want to be on the market of aviation. And they created Airbus with exactly

this type of funding and we've seen that it's a company, it's a successful company, right? So Europe can do it. mean, you're the member states, they put some like agreement on the table, it can happen. So the future will tell. I know many people in Europe that are supporting of this last funding, but you can imagine the administration behind it, mean, the administration, all the admin needed to put such things in place. But that would be something I would...

definitely encourage and it would be good for the market to have more competition.

Billy Riggs (24:29)
Let me ask Vipul something based on what you said. mean, is there a scenario where, you know, Europe becomes so squeezed where Europe has more of an incentive to collaborate with China as opposed to the US and Vipul, how radical is that that idea?

Vipul Vyas (24:46)
In a multipolar world, it's not radical at all. In fact, it's probably the smart move to hedge some bets just no matter how big your hedge is I think the Europeans are probably going to be more open to Chinese EVs than the Americans for a variety of reasons Maybe and then you know, especially if China continues on the path that they

demonstrated with DeepSeek of open sourcing a lot of what they're developing. And China's doing the open source path because the US has not been doing that, though proclamations are doing so. I Lama is one of the few big US bets that is open source. regardless, the point is that Chinese have...

leaned into open source when it comes to AI, that's because that way they can have soft power influence all over the world and the Europeans are going to find that to be very attractive. would think anyone would. Maybe when the US are using TPC. And so I think as just an example, so I think this is going to be a model. And so it'd be only smart to hedge bets. Now, I think from the geopolitical

gyrations are going to settle down by 2027, mostly because the shock and awe is an attempt to get all this done before the 2026 midterms. And so I think memories are going to fade with regard to this unpleasant period. With that being said, a few, that's a few.

Henriette Cornet (26:15)
Thank

Vipul Vyas (26:19)
paths are going to be taken that are going to be long standing and durable, is China's leaning on open source AI to compete with the US, that being attractive to everyone, including Europeans and people adopting lot of that technology. So that's a long-winded answer to your question.

Billy Riggs (26:37)
No, think it's,

yeah, it's exactly what I was thinking as well. I mean, I think when you think about open exchange of ideas, irrespective from what we talked about in the past episode, guarding the gates, great episode, check it out. When we talk about security and we talk about, you know, international exchange of ideas, it does accelerate our economies. And strategically, if the U.S. is going to, you know, start

hoarding ideas, hoarding commercial enterprise from the rest of the world, it does behoove European countries to start trading with Asian countries. So I think, you know, that would be a strategic move on behalf of any other foreign entity to begin free exchange with another polar entity. And that could accelerate their edge.

We say, I didn't always say that to say the best thing for the American edge, the best thing for global competitiveness is for open competition, open exchange. And I think that's what, that's the spirit of everything we've been talking about with automation, innovation, and the spirit of what we've been about is acceleration of ideas and innovation as being something that is about open exchange and open innovation.

Vipul Vyas (28:01)
Well, I think to say one last thing on this topic on my end is, and you again, you guys would know well more than I would, but the other big driver here is military technology. The Europeans and very few people want to be dependent on the US.

Henriette Cornet (28:07)
Thank ⁓

Vipul Vyas (28:18)
for military technology. Everyone's kind of bet on the F-35 and even the Swiss have bizarrely. But there's probably some notional regret. And from what we're seeing in Ukraine, there's a lot of drone technology. Drones and autonomous anything. Drones are just autonomous vehicles of a different sort. They're not for transport, they're for destruction, but they are autonomous delivery vehicles of lethal payloads. And so I think the Europeans

are going to develop their own technology in this this realm because they'll feel compelled that they have to and as a result a lot of that technology has spillover benefits into the civilian space so I think by virtue of the increase of know what the doubling of military spending through a lot of Europe

By virtue of that, there's going to be a lot more tech that is homegrown, not just China versus US, but European developed.

and designed and conceived and built in Europe. I think that people underestimate, there's sort of an assumption that Europe's in the state of malaise. Yeah, there's some truth to that, but I think given the existential crisis that's looming or perceived to exist to Europe's east, I think that's going to motivate a lot. And there's a lot of spending, almost just by sheer luck. If you spend that much money, something's going to get produced that's of use.

and value when I'm talking about military budgets and everyone knows that military budgets are going to be more allocated towards drones and autonomous fighting vehicles and that's gonna like I said spill over to the civilian side.

Billy Riggs (29:55)
Well maybe slightly more controversial, while Henriette was flying west back from Hamburg, I was sitting in London at an AI conference, but something was happening in Austin

and I thought we should maybe talk about this because. we've talked about it a little bit in the media as a part of our our commentary at Autonomous Vehicles and the City Initiative

but Dr. Cornet, we haven't had a chance to really debrief on What are your thoughts about what's happening in Austin right now with the Tesla robotaxi experiment?

Henriette Cornet (30:31)
Yeah, that's, I don't want say that's a topic I wanted to avoid, but when you say, should we talk about it? I was thinking, should we talk about it? in terms of, that, I don't want to say relevant because yeah, of course it's relevant. That's it's very relevant, but it's a bit of a pain point in this entire discussion because yeah, I had my vision of having on one side, the company's targeting high level of automation in restricted area.

in cities like Geofence, like Waymo deploying in more and more cities, but directly with a super high level of automation. And on the other side, having players more like the automotive manufacturers driving everywhere, I should say, but with lower level of automations. And this vision, like I could put like every players nicely in these little boxes. to say, and Tesla is kind of mixing all that.

because they claim they can drive everywhere and they claim they have high level of automation. So it's a bit, so we need to look at that closer if it's really the case. And it happens that as you like saying autopilot is not self-driving, I think it's entirely correct. It's not, it's a level two, maybe a bit level two plus plus like they like to mention and same for your audience.

There could be an interesting link to the different level of automation to be described there. And what they're doing in Austin now, it's still not really, it doesn't look like full automation because they have the safety driver, because they do have, apparently they have a car behind following the taxi that is driving. So it sounds more like a pilot, which I love pilots. mean, I'm really in favor of pilots. It's how you learn, right?

And Austin is a fantastic place to start. Mainly, I think it was also because of regulation there and it was easy to put vehicle on the road. But apparently recently there have been also changes from Texas to say they will, they may put more rules to get a permit to deploy autonomous vehicles. And I think that's a consequence of Tesla maybe deploying too fast, too quick without having the technology completely.

safely provided. So for me, it's as much as I like action and as much as I like having vehicles on the road, it may create some confusion to call what Tesla is doing really a robot taxi or self-driving. For me, it's still the area of ADAS, which is more like supporting drivers in terms of safety and comfort.

rather than really going without anybody, like without a driving wheel or nobody behind the wheel like that. I don't think they are there yet. And I'm a bit skeptical how much it can impact the entire industry of L4, having people confused about that. I think people will not make the difference between a Waymo and a Tesla. And that's kind of problematic. did, by the way, I have to advertise my own work there that I did a video.

comparing the two like Waymo and Tesla and how they differ and could be interesting also for your audience to check to make sure that there is no direct association being made. The approaches are so different.

Billy Riggs (33:42)
Yeah, that's Urban Innovate Talks, right? We can put a link for that in the comments for this. But also, I think I want to pinpoint ADAS for all you folks that aren't transportation speak nerds. We try not to do too many acronyms here. It's Advanced Driver Assist Systems.

Henriette Cornet (33:44)
Yes. Thanks. In the comments.

I'm sorry for that.

control, all this aspect, starting with parking assist. These things will gradually evolve, like all the time. You can expect really your car to drive by itself on highways pretty soon. I'm sure all the automotive manufacturers work on that. But does it really mean it can go into the city, interact with pedestrians and cover all the situation? There is still a long, long stretch. And that may not be clear always in the head of everybody.

Billy Riggs (34:27)
And I think one of the things that I'll dispel now is that if you're watching the news, and if you're sitting with your family or your kids and they're saying I saw in the news that Elon Musk delivered a Model Y. Somebody purchased a Model Y and he delivered it to their house. And that's really cool. There's probably a lot in there

behind the scenes that you're not seeing. So I think we are actively skeptical. But I actually want to turn it over to Vipul because I was at an AI and ML conference and doing kind of refreshing my own programming and

technical skills this past week. And I do have I'm going to say I have a lot of respect for the Tesla approach to self driving, because it is approach that says, hey, this is an approach where we can potentially have a very bold approach. I just don't think they can get there as fast as they say they can. I think it's probably a 10 year approach to self driving. And I'm going to say that because

The traditional approach to self-driving has been the, the, what was called the DARPA challenge. And it was the defense, more of a robotics approach. There was a, defense department had what was called the DARPA challenge. And basically all these, there's nerdy MIT roboticists. They built these robots and these robots got into this arena and they would either race or they'd fight each other. And there, these were these computer scientists and roboticists and they'd, go in there and these robots would duel one another.

And for a long time, was just these nerds doing this stuff. And then they decided they'd all spin off and buy the, know, build these companies. you know, Kyle Vogt was one of them who ended up and started started Cruise. Kevin Peterson, who ended up and started Marble Robot and then was one of the original people that actually worked on the trucking stuff for for Waymo. And now he's got his own robot company.

Chris Urmson who's now CEO of Aurora was also part of that whole crew so you can see a lot of these people who originally part of this DARPA crew were the traditional roboticists. They didn't rely as much on software. They relied a lot on hardware, but then you get Elon Musk here that's like, I can do all this with software and basically skinny skinny skinny the hardware. Just do it all with cameras and do it all with cheap cheap cheap cheap cheap because I'm doing it with software.

Now, is it possible? We don't know. Vipul. Can you do all this with just like the human mind of like a Turing test? Can you basically say software can function like a human mind? Humans basically have more or less one sense when they drive a car. Their eyes. More or less, right? Some part their ears, but more or less their eyes.

So can a camera, can a camera function if basically a nominal touring test, a nominal, you know, this, this idea of a sophisticated AI long-term. So is this Tesla experiment long-term? Could it work? I would say my guess is probably, but necessarily in the next two months, I would say no.

And this is why I think it's really interesting to this question. And that's why, I don't question the approach. I just don't think it's possible in the near term.

Vipul Vyas (37:50)
I don't know first, I think that what they've done is they're moving towards this sort of unsupervised learning world where, you in the past it's sort of like, okay, you gotta train the AI on a stop sign. What does that mean? How do you interpret it? Have you?

react to seeing that how do you recognize it or what do do when you do recognize it. They've gone into let's just process all the driving data that we have. Let's just see what happens empirically when you have all these cameras on existing Teslas.

If you had access to some of that data, you can say, here's how many millions of driven miles there are and what behavior there was in reality and do less supervised learning. And theoretically, that approach lets you learn more faster. ⁓ And so that's going to be the same kind of thinking that's going to go into Optimus, that humanoid robot.

Billy Riggs (38:45)
Yeah, they don't pre-map either.

Vipul Vyas (38:53)
So this is not just for the purposes of autonomous vehicles, it's for autonomous anything. And that's how humans learn. don't, to some degree, you have school, of course, which is much more rigid, but humans evolutionarily learn through lived experience. And that's essentially what this is, is lived experience. And what is different about human lived experience and AI lived experience is that...

My lived experience is my own and I really can't easily tap into someone else's unless I'm invested in reading someone's biography and even then that's very limited. But, you know, these machines can absorb the lived experience of hundreds of things, hundreds of millions of people, millions of other objects for that matter. Like another Tesla and what it experienced through what it captured on its cameras and whatnot.

And so ultimately that is probably a better long-term bet in terms of how to train these things. But to your point, it's not going to happen overnight, but it's probably going to be one of these things, you know, very slow and then all of a sudden. ⁓

Henriette Cornet (39:54)
Yep.

But where we would intervene here is really at what cost and when I mean cost, don't mean dollars. mean, cost of people's life. That's what scared me with this approach. So doing trial and error where we are talking about road safety. I don't know. have big, big, big limitation on that, on that kind of approach.

Vipul Vyas (40:13)
You

have the same thing with the humanoid robot. What will it do to someone? I think there's the same risks.

Henriette Cornet (40:20)
Yep.

Billy Riggs (40:21)
Yeah, I mean, we should all go watch the unbridled, you know, murderbot series, right? You know, what is Alexander Skarsgård going to do in the next episode? We don't know.

Vipul Vyas (40:25)
All right,

And I do have

to go, but I think it's an interesting, this is a big subject. I think the other big thing with AV is just what do political leaders need to be thinking about in terms of the confluence of AI, transportation, built environment, electric vehicles, all those things, the convergence of those trends.

has massive implications that no one's really, I don't think thought through very well. ⁓ There's a huge business in helping people think through that, think, but they just haven't. There's a lot of implications and paths that can be taken. And I think not doing anything is gonna be a mistake.

Billy Riggs (41:05)
Yeah.

Henriette Cornet (41:06)
him.

fully agree with that. think we need, I like what you mentioned with this poverty of aspiration that exists here and there, unfortunately. And I really wish, and mainly I see that even more, I'm sorry, but I see that a bit more in the US having such a gap between the industry and the public sector in general. And my field, know, like public transit, public transit authority where there are some actors

spread over the US, mainly in like rather big cities that are really like, you can see they have this innovators mindset and they want to know what's, they want to benchmark the market. They want to know what's out there. When can they have things on the road, et cetera. But too many transit agencies are reluctant to innovate. And that's, that's, yeah, that's the worst because they will just disappear. I mean, I think at some point, public transit will just disappear if you don't have this kind of looking forward mindset.

Vipul Vyas (42:12)
Yeah. Yeah.

Henriette Cornet (42:15)
You know, people, you mentioned all this aspect of AI and everything that is happening, et cetera. But at the end, we are talking about quality of life, right? Like, what do we want in our city? How do we want to move around? How do we want people to have access to job, access to health, access to leisure activities and like this vibrancy that we have in so many cities? How do we want to keep that? And not everybody being completely, you know, the worst model for me would be we all, we're all at home watching Netflix, like a video on demand all day.

and then we take our autonomous pod on our own somewhere else and we like there is no social connection anymore.

Vipul Vyas (42:51)
That

is the default state. And the reason that's the default state is because what you described is the world.

that extracts the most wealth from the individual.

That's the most efficient way to have a person part with their value. Right? Is this sticking in front of a TV for as long as possible? They can eat bad stuff. Yeah.

Henriette Cornet (43:06)
Yeah.

consuming ads.

Billy Riggs (43:13)
it sounds

like that other that other movie ⁓ no no the ⁓ the one of idiotic see yet ⁓ it's exactly idiotic see ⁓

Vipul Vyas (43:18)
Wally?

Idiocacy

Yeah, mean, there's a reason. It's a very extractive way of heading

up the world, but that's the way the world is default headed towards, unless someone consciously makes a decision not to do that.

Henriette Cornet (43:37)
Yeah, and I hope the questions we have would help a bit rethink of what we really want, so to say.

Billy Riggs (43:38)
but I can also

But I think maybe I can summarize where we are too and just say, I think it's not just about, it's not just about automation, it's about kind of intelligence. And so as we're thinking about automation, it's about how we manage these fleets. It's about predictive maintenance, dynamic routing. It's about orchestration. And, but it's also about activating people, getting them out of the house.

enlivening their lives, increasing livability and making places, better and more livable connecting humans, and reminding us that, we're not robots. and I think if I was providing advice to city leaders and maybe this can be maybe our final word is it, I think what I heard from you all is that.

the worst action is inaction and perhaps it's okay to be approximately right than precisely wrong.

Vipul Vyas (44:35)
but I think to Andrea's last point on Netflix the former CEO of Netflix once said that his biggest competition was sleep speaks to the fact that without action without intervention without planning we're gonna head to a world that no one wants to really be in or part of

Henriette Cornet (44:54)
No, think there is hope. think there will be a moment, I'm kind of expecting a moment where people will be fed up with AI and their smartphone and their screen and they would like to...

to gather again and maybe it would be more on community level and not really on this Megalopolis kind of mindset. But we will see that and there we have to think of what's the best way to move around and if it's automation or not. Vipul I love that you mentioned also the "radical" way of moving which is walking. I think that's also something we should keep in our discussion, active mobility, etcetera. yeah, thinking of mobility as something holistic.

that can make city more livable.

Billy Riggs (45:36)
That's right. Vipul, anything else?

Vipul Vyas (45:39)
That's it.

Billy Riggs (45:41)
Well, I hope you learned something. Thank you, Henriette, for joining us. You're always welcome. ⁓ Yeah, and thanks everybody for listening. Thank you for subscribing and help us make cities better. Stay tuned, subscribe, and join us next time.

Henriette Cornet (45:43)
I do them. Thank you so much, guys.

Join our newsletter

checkmark Got it. You're on the list!
University of San Francisco, Autonomous Vehicles and the City Initiative. All rights reserved. 2025