The Vergecast - Argo’s CEO explains why its fleet of self-driving taxis won’t be all-electric (at first)

Episode Date: February 25, 2020

One of the burning questions facing the world of self-driving cars is whether it makes sense to go all-electric or not. Some, like GM-owned Cruise, is all in on battery-electric vehicles. Others are going half-and-half, like Waymo building a fleet that includes both all-electric Jaguar I-Pace SUVs and gas-burning Chrysler Pacifica minivans. Argo, the Pittsburgh-based self-driving company backed by Ford and Volkswagen, has concerns about an all-electric fleet, especially when it comes to the need to recoup the cost of all the expensive technology that makes the car autonomous. The company’s CEO, Brian Salesky, sat down with The Verge editor-in-chief Nilay Patel and senior reporter Andrew Hawkins to explain why. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Support for the show comes from Retool. Too many companies run critical operations on duct-taped spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Prompts something like,

Build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data, in your cloud with enterprise security built in. Go to retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all. I'm Skyler Diggins, seven-time WMBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years covering the biggest names and stories in sports and mom. And this is Am Mom, a community for athletes, game changers, and moms of all kinds.

dropping May 14th. Tap in with us. Hey everybody. It's tonight from the Vergecast on this week's interview episode. Andy Hawkins and I sat down with Brian Seleski. He's a CEO of Argo AI. Argo's a self-driving car company. They've taken a huge investment from Ford.

Brian's got interesting ideas about how self-driving technology will actually take shape over time, how it will be integrated into cars. We talked about the pitfalls of self-driving. What he's really seeing is the challenges of the industry. And, of course, I demanded to know when it's going to happen because everybody has a different answer to that question. Brian was really interesting conversation.

He knows a lot about this field. Check it out. Brian Celestky, you know the sea of argo.a.i. Do you call it Argo or Argo.a. We'll say Argo, Argo, AI. Either is fine. You never use the dot. Rarely, unless we're pointing someone to our website. By the way, I've got Andy Hawkins or senior transportation editor here. Hey, yo.

So tell me about Argo. You make self-driving cars? Well, we do. We make the software, the sensors, computers that go onto a car, that allow it to be self-driving. And we're partnered with a couple auto companies that do the vehicle part. So I ask every self-driving executive that comes on our show of this question. Our self-driving car is going to happen.

Yes. When? It's ready when it's ready. That's actually the best answer. It's the best non-answer answer I've got. Most people are like, it will. I promise. Hazy.

I declined to answer when. No, I mean, look, I've been doing this for well over 15 years now. And it moved probably, I don't know, maybe five or six years ago. moved from an if to a when thing, right? We spent a lot of our time doing assistive technology. How can we aid the driver? How can we augment the driver in a number of different projects back in the day? But when it came to driverless cars, I think it moved from if to when for me about five years ago. So you were previously at Google and you left to do this one. Yeah. What was the thing that made you think, this needs to be a different company? Well, I left to, first off, it's always been on my

list to start a company someday. I just didn't know when the time might be right. But the funding environment was right. I felt like there was a way to build a company that was strategically aligned with an automaker to in order to scale it out. Prior to Argo, I've worked most of my career working on things that are in numbers of hundreds. And I want to get to, I want to build technology that will be out in the million someday. In order to do that, you know, making building, manufacturing a vehicle at scale is not easy for any startup. There are companies doing. it that have been, that have found some success. Don't get me wrong. But I wanted to stick to what we know, which is the robotics part, and then partner with the car companies to do the car part.

Argo's introduction into sort of the public consciousness was really interesting because very few people had heard about you guys until Ford made this announcement that they would be investing a billion dollars. I was always the guy behind the scenes, right? But I think the industry knew who I was. I'd like to think that. I mean, I had banked a number of years at that point. For sure. But, you know, at Google, there were very few people. people sort of at the front line. And frankly, while I was there, I was quite happy to be the guy behind the scenes. It was a great place to work. But what was that like, sort of having that sort of be sort of your broader public introduction, which the announcement of this very

major investment from obviously a storied car company like Ford and having that sort of like billion dollars sort of being the thing that kind of hung over the announcement, too? I think it was helpful, at least from a recruiting standpoint, because it helped our employees see a little bit about the mission and how we were intending to go to market and how we, that we were intending to partner. It certainly showed other car companies that there was a willingness for us to kind of work with their processes to do things again at scale and do them safely. It's been really helpful to us that we have been able to be heads down with some stable funding to just go make progress and build the product. It's really hard to build a company when you're simultaneously fundraising, selling, and sort of building something from scratch. And I have seen that part as well in other things that I've invested in in the past. And it's like it is hard.

So that funding was very helpful to us to be able to make progress very, very quickly. What's different about Argos tech than things like people are familiar with, like Waymo or Cruz or something like that? Well, I think our approach is, I usually leave the comparisons to you guys, but our approach is to focus on the urban cores. So it tends to be a little bit lower speed, but a very high dimension of complexity. So we operate in South Beach. In fact, I was just there and, you know, the car operates super well, even with the total influx of just pedestrians everywhere. Forget crosswalks. Crosswalks don't matter. Everybody's going to the beach and they don't care. if there are cars in the way, right? And so, look, we want to go where the people are. We want to go where people want to be because we know there's lots of trips. There's lots of demand, and there's a good business case to wrap around that. So we've traded off to, we've focused, we're focusing on those really high, complex areas, lots of interactions per mile per hour, however you want to dimension it. Is that a compute problem or a sensor problem? I think we tend

to, just for the verge, we tend to focus on the hardware. What's bolted the outside of the car so you can see the world. But are you more focused on? on the what am I seeing and how do I get around it or you we have to detect everything first. You know, robotics is a funny thing. It's weak. We called a tightly coupled system, meaning that everything has to work for you to have a good end result. If any one piece of the picture, any one piece of the system is performing suboptimally, the thing doesn't work. That's the thing about robotics. So it's all important. We need certainly a lot of compute. We certainly need sensors that give us 360 degrees of awareness around the car.

We're a believer that you should use multiple modes because different sensor types fail when compared to others. So the strengths of one complement the weaknesses in others. So we use camera, radar, and LIDAR. And when you're working in a very complex scene, it's not good enough to just say, and when I say complex scene, I mean lots of cyclists, lots of pedestrians, lots of cars that are everywhere, not following necessarily the rules of the road. In order to interpret that and navigate it, you really have to have a very, very, very, very acute and precise sense of what is what in the scene. It's not good enough to say, hmm, that blob of like points from a LIDAR, we think that's a pedestrian. No, you have to know

that that's actually a pedestrian climbing out of the vehicle and they're most likely going to go around in a particular path and you want to give them plenty of margin and room if it's safe to pass. So the technical term might be seen understanding. That is super, super important. We hear about some of the scenarios that self-driving cars have figured out, whether it's traffic signals, certain intersections. What are some of the scenarios that are still proving to be a challenge, would you say, for cars to figure out? Does it sort of go into the area of edge cases, or there's still just sort of very, like, basic things that still need a lot of work done to it? So the broad strokes answer is it's things like really bad weather. So if you have,

and by the way, if there's snow in the ground, that's not so much a problem. It's more the falling precipitation. So if there's falling rain, falling snow, fog, those are difficult problems. And before you even get to the software part of the problem, you have to get past the fact that the sensors just see the world fundamentally differently. And so we have to build just new models and new ways of tackling that. Maybe there are new sensor types we could look at. But it's going to take invention, I believe it will take invention both at the hardware level and the software level to tackle that problem. You mentioned edge cases. You know, edge cases are a funny thing. As a human driver you put on however many miles a week, or maybe you don't drive, but a typical person in their

commute will put on so many miles a week. And they might see one or two anomalous things. If you have a fleet of cars every day putting thousands of miles on, all of a sudden what you thought was infrequent actually turns out is really frequent, right? And so that is the other challenge. And yeah, absolutely, we have to build a system that's resilient enough to handle all of these pretty frequent things that occur that to you may seem abnormal, that actually we're learning is pretty normal, actually. So let's take a step back and talk about why you guys are doing this. You know, why do we need self-driving cars and what's sort of like the business case that you see for the technology you guys are working on? So I think the answer that you probably hear a lot is about around safety and that's super

important and we believe the same. We believe the same. Human drivers are not great at driving. They're distracted. Many of them. There are good drivers out there, but there's all So there is also a large distribution of drivers who I think would prefer not to be driving. So we want to not remove driving. We want to augment it and give people another choice. And a lot of people don't use rideshare services today because they want a more personal experience. They want to listen to the music they want to listen to. Maybe they have to carry things around with them.

Maybe they have whatever it is. They want a more private experience, a more tailored experience. Or maybe they're afraid of their driver. They're in a situation. Well, you know, for a lot of people, especially for women, for women writers, it can be a, you know, a crapshoot in terms of what kind of driver you can. And I am having a really bad run of rideshare drivers right now. Like, I just am. It's not good. We could talk about that later, probably. So the safety is number one. And that is absolutely a mission that drives the company, our employees. It's why I'm involved in it. I think what doesn't get talked about enough, though, is also the problem that cities experience and how this can be. a solution to at least a dimension of the problems that cities have today. Cities have unprecedented congestion.

There's more cars on the road than ever. The system is in gridlock. People's ETA, the ability to get to where they want to go quickly is increasing. Average commutes are increasing in time. And so how do we solve that? The good news is that the more autonomous vehicles that are on the road that are deployed in a shared context, in other words, you don't own it, you just use it when you need it. It means that the vehicles can pre-position themselves at night when there's less congestion on the grid for where the demand will be in the morning.

It means that we'll be able to, after it drops you off, can go on and serve the next person. It won't be hunting for parking or consuming real estate in the city. Rand did a great report that showed that up to a third of the real estate and a city core is devoted to parking, which is a shocking statistic. Imagine if you repurpose that for other things that people want, whether it be parks or affordable housing, whatever it might be. But that's going to take a long time, right? It's going to take a long time, and I'll get to that part. The last thing I will, but the other third thing I want to get into that I think is important is that in addition to changing the landscape of cities, we can also tell vehicles, an autonomous

vehicle will listen to a, think of it as like air traffic control for cars, right, that will more intelligently route it so that we can alleviate and deal with the congestion. Maybe I add 30 seconds, not much to your estimated time of arrival, but I've saved the other person 10 minutes because they need to take that route you didn't. So we can load balance how the cars use the available road space. Now, we're not solving that problem at Argo, to be clear, but we're going to see companies that are also lorking in that space. You can't necessarily tell your Uber driver to take this particular route, but you can tell an autonomous vehicle if we all work as part of a shared system for the greater good. So I think that's some of the

huge advantages. Now, none of this stuff happens without having sort of a more of a penetration of autonomous vehicles in the, in the global, in the fleet, or I shouldn't say global, but at least city by city. And so, yes, it's going to take many years, but we've got to start now if we want to get there. I'm having difficulty sort of squaring the notion in my brain that the solution to all sort of the problems that plague cities, whether it's traffic congestion or infrastructure problems, is more cars. I feel like there's a growing sort of number of people who say actually the solution is less cars. And we need to start thinking about policies and solutions that sort of redirect

cars out of cities and try to find ways to get people out of, whether it's right hailing or taxis or what, or personally on vehicles and not to public transportation. Obviously, these are huge issues and not something that Argo is necessarily geared towards solving. But how do you sort of address that issue? You're totally right. So more vehicles is not the answer. Again, encouraging the use of shared fleets helps because we then can create incentives to not bring a personally owned vehicle into sort of this, at least the city center where there's the most congestion and parking issues. The other thing that we can do is, or the other thing that we have to realize is when I talk to city leaders, yes, I would love to have huge rail systems that would take

us anywhere we want to go. Yes, I would love to have a subway in every major city. Yes, I would love to create more bike lanes and take away road space to give people other modes of transportation. I would love to save the public transit systems that tend to be bankrupted in a lot of cities for a variety of reasons. But we can't solve all of that at once. And the fact of the matter is when we talk to city leaders, they're stretched thin just doing the simple things, let alone figuring out how to put an elevated rail system in, for example. And so my view is that while we need to keep the road infrastructure running, let's find a way to reduce the number of vehicles. Maybe we can look at higher occupancy autonomous vehicles over time, and we can use that road space more

efficiently. Because I don't think any of us have the silver bullet that's going to solve the funding problem that exists that would otherwise allow us to live in a car-free world and still get to where we need to go. So when you're talking about reducing congestion, and you're talking about maybe once we have a fleet of sufficient size, we can start programming cars to go in different ways to reduce overall congestion. That either implies that you will operate all the cars, which would be great. And we're not doing that. Right. That's not the suggestion. I'm assuming you and your investors would be happy with that outcome. There's only one company it's Argo that got all the cars. Or you've got to be interoperable with every other kind of

self-driving car. Or there needs to be some sort of statewide federal vehicle to infrastructure communication system that exists. Which of those are you betting on? Which of those is going well, which is going poorly. So I don't know, it's a great question. I don't know how all of that's going to play out. We are not building the kind of the traffic cop, so to speak. And I know there are other companies out there that are very nascent that have this idea, this ambition. But for those companies to be successful, they're going to need to be willing to really work closely with the cities and engage.

Because they're going to have an unprecedented amount of data about the movement of objects in that city. I use objects purposefully if it's conceptual. they're going to know a lot about everything and how it moves. And so there's really important issues here that they're going to need to sort of work through. I still believe, though, that without that kind of traffic control system, I think it's going to be very difficult in order to deal with the rising congestion issue. I mean, we're starting to see some of this play out, but in a different space, in the scooter space right now.

You got all these scooter companies that want to drop their vehicles into cities. and, you know, at first that seemed to catch cities off guard, and now cities are sort of playing catch up, and you're starting to see some of them, like Los Angeles, play really aggressively in terms of data sharing and, you know, mandates on the companies. Do you see that as sort of being sort of like the model that's going to be carrying off into the future for AVs? Yeah, like I said, I don't know what the right model is, and I'm certainly not able to sort of lay out the master plan. And I would say that Argo and our partners were open and we're listening and we want to tap into some sort of standard that would allow this to happen. It's for everybody's greater good.

We have a vested interest in getting customers to where they need to go as efficiently as possible. So if there's a better way to do that, we're going to make use of it. This is all like the second order problem. This is like once we win, we can foresee this set of problems coming to the fore. Like how are all the cars going to talk to each other? who's going to control the data, how we keep a private. But first, the car's got to drive itself. It does.

And so I guess my big question is, you've got safety drivers in your cars right now? We do. Presumably. How long until that driver gets out of the car? It's ready when it's ready. You know, we look at a lot of the data internally, whether it be events that come in, we look at all our testing, we look at those results on a daily basis. We're constantly improving the system.

You know, the answer is I don't think it's entirely up to us. Even if I thought, hey, we're going to be ready on this date later this year, even if I said that, I think that cities also want to have a say. They want to understand. And I can't exactly time how long it's going to take in order to get, you know, mayor, XYZ up to speed and for the community to be excited for us to be there. Last thing I want to do is build a company that launches a bunch of technology that consumers don't understand or are afraid of. We want to build that trust, and that's only going to happen over a long period of time. So I don't think that there's a binary switch that gets thrown where a company puts a bunch of vehicles on the road, and then boom, everything's autonomous. There's no driver in the car. I think it's a staged thing, right? We have two drivers in the car today.

As things mature, I'm sure we'll move to a single driver. After that, I'm sure we'll move to a point where maybe there's a person in the car, but the person's just there to educate the riders, answer their questions, and get them comfortable. This is a long-game thing, and it's going to be a multi-phased approach. It's been kind of a momentous year for autonomous vehicles. We've had some of your competitors showing off their wares, basically. You've got Waymo doing rides with nobody in the front seat. You've got Cruz just recently unveiling its first purpose-built vehicle that they say is going to go into production. Do you guys see value in sort of those kind of like flashy public demonstrations is like, hey, come look at our tech.

It works. Come look at this car that we're going to make. It's going to be so cool. you guys seem to be sort of playing things a little bit closer to your vest. Is that in purpose? Well, I think everyone's in different stages of development. And for us, we want to share things to the extent that it, going back to the education topic, we want to share things that helps increase the awareness level of communities.

I'm sure that the folks in Chandler or wherever Waymo is operating with their first driverless system, I'm sure that they have learned a lot through the engagement with Waymo. and they've got a plan as to how they launch and they're working with the city and those are great things. We need to, as an industry, we're all going to be building up our playbook as to how we launched in each city

and it's going to be potentially different city to city depending on what the community wants. And that's the thing that we talk a lot about at Argo is that this is a city by city, street by street thing. There's no silver bullet that launches this everywhere in 24 hours. It's not like an app we put on the store. on the app store that everybody can just go download the minute you're ready to launch.

It's a very different type of launch than is typical for most industries. I mean, that presents enormous challenges, I would imagine, especially if you want to, like, scale a business. And that's something that we hear from Ford, one of your big backers a lot, is, you know, we have to build this business. And, you know, the whole intention is to launch at scale. At least that's what the company said in terms of when there's a, you know, when the technology is ready. Street by street, block by block, how does that sort of what sort of tensions does that create in terms of wanting to scale a business? Well, when we say scale, we're talking about, I think scale word is used maybe too generally sometimes.

Scale for us means that we're building a system that can go to multiple cities without having to completely redesign it. There's some technical approaches that you could take that would allow the vehicle to work perfectly in one city and then not be able to take it to the next five or ten. So Ford has obviously the manufacturing scale, which is important, but they also. also realized that at the same time, this is a brand new technology that people need to get familiar with first. And so our plan is that we're operating across many cities in what we call kind of a beachhead. We establish an operating area and then we slowly grow it over time, but we do that in parallel. So we're operating in D.C. We're operating in Miami, Austin. We've got a few more

to follow. And this allows us to do the mapping, to engage with community leaders. We build a great, excellent workforce in these cities that is helping us to test this technology, who have local knowledge that really helps us develop the playbook I was talking about earlier. And so my viewpoint is we are scaling by doing that. Yes, it's small and no, the ramp is not maybe as steep as what everybody would want to project, but we're getting learnings every day on how the software works across many, many cities. And that's super powerful. Do you think that autonomous vehicles are in this troth of disillusionment right now? Well, I don't know where we are in the hype cycle.

I do think that the overall media engine has maybe in the last few years increased or maybe set expectations where a lot of what we just talked about was somehow not accurate. It's going to happen. Brian, you're just too much of a downer. Tomorrow, we are going to throw the switch. And I just don't believe that's the case. And I think it's starting to become true. And I think a lot of the writings that we've done where people said, gee, you're just, you know, you're not, it's going to happen faster.

I think people are starting to realize that, no, it is going to be gradual. But it is also, I'm also very optimistic about the future. Again, the safety benefits, the way in which cities can change over time is super powerful. And so we're a huge believer in the tech. It's just going to take a little bit of time. And so I don't think it's the trough of disillusionment. We've always approached this with sort of a realistic set of expectations. So you've got a different kind of model and attitude about this than some of the other folks that we've talked to.

And I just want to push on a little bit. You know, you're talking, again, kind of at the second order of things, right? The long goal, the long game that you're playing is eventually cities will change around self-driving tech. And we're going to build technology that enables that and lives in that world. Right now, I'm just looking at your website, that looks like a Ford Fusion with a, a hat on it. Right? But eventually that's off

we get smaller. Ford will redesign the car around your stuff. That's a prototype. They're actually building a purpose-built vehicle. They haven't announced it yet. There you go.

What's it called? When does it come out? Well, I can't. I can't give you answers to any of those because they haven't announced it yet. That's what I'm signing up for. Are you trying to get me in trouble with that? That's all I'm here to do.

I think everyone sees that. Next to that, though, you have a bunch of carmakers who are slowly but surely iterating on their assistive technologies, right? And that's the other path they're all taking. You've got GM doing Super Cruise and the Cadillacs. It's getting better. Mercedes has a version. Tesla just straight up calls it autopilot, right? They're just, they're taking basically cruise control saying it's a little

bit better than before. It can keep you in your lane. This is how we're going to march forward. That is a pretty familiar pattern in consumer tech, right? Like here's this, here's the inkling of a, feature or capability, and we're just going to iterate until it becomes the whole thing you want. Why are you betting on sort of the big jump as opposed to the iterative approach? Well, I think there's a time and place for both. The iterative approach makes drivers better today and prevents drivers from maybe making bad decisions. That's excellent. We should feel that as fast as we possibly can.

That wasn't the way in which Argo was founded was not to solve those set of problems, but there's plenty of companies that are working on it. And I 100% support all of that. I would like to see that stuff on every single car. fast as humanly possible. Automatic emergency braking is can be and has been a lifesaver. We need it on more cars. I'm with you. I'm with you entirely,

right? But if we want to talk about you've used the phrase second order a couple of times, second order to me sounds like it's maybe a lesser impactful thing than the first order thing. I don't think that's the case. Autonomous vehicles has the potential and will be realized someday. I'm struggling because I don't think any of us knows how to put a time on it, but it has the

potential to completely reimagine how cities are built. It can plug into a smart system to address congestion issues. It will allow people to get around cities much easier. It can, it has the ability to serve areas that for whatever reason, there's lots of reasons, but for whatever reason might be underserved and where the last mile issues are really troublesome. That's super impactful stuff. And that's, those are first order level, that's a first order level of impact. And I'm not saying that because we're working on that, we should just wait and just deal with the vehicles we have today. We should be improving the whole ecosystem. And to your point, there's many years before all of this gets rolled out truly at scale across hundreds of cities and, you know, millions of people.

In the meantime, we absolutely should be making cars safer. Yeah, I guess the reason I keep saying first and second order. I'm glad you're pushing back on it. It's the second order set of challenges. Like, once you have the self-driving car, it becomes very easy for me to orient everybody that I talk to our audience to, now the city is going to manage a bunch of traffic data and send instructions of perfect compliance to cars in a way that we know that human drivers just won't listen because they won't. I won't. But the robots might, we hope. That's why I keep saying it's second order. The first one, it actually, you're talking about priorities then in terms of what? Yeah. I think I'm really getting at here is it seems like you're betting on the paradigm shift.

And everyone else is sort of like, how do we, how we inch our way closer to? I'm saying it's inevitable. And it's going to take many years. We've got to start now. But we also have to be doing things in parallel. We should be working on making every type of vehicle and mode of transportation. Doesn't even have necessarily be a vehicle. We should be making every mode of transportation as safe as we possibly can.

Because changes in transportation changes to how vehicles are fundamentally architected and designed. It just takes years for it. pull-off. There's a whole supply chain that needs to be built. There's validation that needs to happen that takes a long time. It's capital-intensive stuff that only pays back over many years. So if we want to see any positive change in a complex area like transportation and automotive, we have to be playing for the long game on all these things. We've also got a climate change problem, and obviously transportation is a major factor in that. I think around a quarter of emissions due to transportation. How do you have a lot of

How do you guys, you guys have not said in terms of we're going to go all electric or, you know, you got the Ford Fusions that you're using now, but have you sort of started to think about in terms of for the fleet, what makes sense? Absolutely. In terms of lowering emissions? Both of our partners are all in on all electric vehicles. They each have solid programs in their cycle plan as to a solid lineup of electric vehicles in their future, right?

Now, if we start talking about how it gets phased in over the near term, say the next five to ten years, the way I look at it is hybrids are a good balance because the technology itself to drive a car consumes quite a bit of power. And this hasn't been really written about a lot, but there's a whole utilization thing that's important to make the business case work. The cars need to be on the road earning money. Okay. The longer it sits at a charger, the more time that goes by that it's not making money. And the potential, if you use fast charging, it's also destroying the battery chemistry, which then has you replacing batteries on a more frequent schedule, which also is not good for the environment. So the way we look at it is hybrid makes sense in the very near term.

Okay. And then as battery technology improves, as we can put more and more battery power, which gives more range onto these vehicles, it can support the technology and we can phase in a fully electric vehicle. So, again, it's going to be a time-phased thing. By the way, this changes, these economics are very complex. The range of a vehicle depends on so many factors. It depends on the type of vehicle, the type of battery technology that's being used and so on. So everybody has to pick something that balances the business case,

but that eventually gets you to that electric future so that we also do the right thing for the planet. Support for the show comes from Framer. Framer is an enterprise-grade, no-code website builder used by teams at companies like Perplexity and Muro to move faster. With real-time collaboration and a robust CMS, with everything you need for great SEO, not to mention advanced analytics that include integrated A-B testing, your designers and marketers are empowered to build and maximize your dot-com from day one.

So whether you want to launch a new site, test a few landing pages, or migrate your full.com, Framer has programs for startups, scale-ups, and large enterprises to make going from idea to live site as easy and fast as possible. Learn how you can get more out of your dot com from a Framer specialist

or get started building for free today at framer.com slash verge for 30% off a Framer pro annual plan. That's framer.com slash verge for 30% off. Framer.com slash verge. Rules and restrictions may apply. Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers.

That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward.

Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. There's a part of me where I'm like, okay, you can drop the tech of self-driving onto an electric drive train, a gas drive train, a diesel drive train, whatever. and you've still got a steering wheel that takes inputs and turns the wheels, and you've still got... Or maybe you don't. Maybe there is no steering wheel. Well, sure, but when you drop it onto one of these cars. Like, you still have to turn the wheels, you still have to go backwards and forwards.

You maybe have to shift gears. But it doesn't matter what's providing the energy of the car. Is that true? Is that just me making a big assumption, or is that... Well, this gets a little bit technical. I mean, every car is unique in terms of its electrical architecture. Whether it's an internal combustible engine, whether it's hybrid or all-electric. there are different nuances here.

Certainly the way in which we tap into the vehicle, there are a couple of things that must be true. There needs to be a fully redundant and fault-tolerant braking and steering system, right? So you need to make sure that on whatever vehicle you add the self-driving system to, that if an electronic controller or think of just as a simple computer that controls the braking system or that controls the steering system, if there's a failure, it needs to continue to be able to break and steer, right? And so the way in which that can get accomplished, there's lots of ways to do that. But those are kind of table stakes as to how you think about retrofitting an existing vehicle. Now, as we go forward, we'll certainly be looking at ways to make that design more efficient

when we look at a completely purpose-built self-driving car that was never meant to have pedals or a steering wheel. But an ice engine or an electric motor is not, that's not changing your design. It's not changing the design substantially. Now, if it's an ice engine, there's not the amount of power available to power the electronics without putting a fairly large and sort of annoying to integrate inverter on the car. Right? You know, back in the old days, if you want to reminisce a little bit, we took, when we did one of those DARPA robot races called the Urban Challenge,

we had a Chevrolet Tahoe. It was not a hybrid vehicle. We actually added an alternator that was fairly sizable onto the engine. But in order for that alternator to provide continuous power to the electronics, We had to run the engine at a higher idle. It blocked a substantial amount of free space under the hood that we thought, gee, wow, it's really convenient. We can stuff an inverter in there. But you know what?

It blocked airflow, which ended up melting wiring harnesses. So we had to put a hood to direct airflow over the fuse box. I'm probably telling you more you want to know. No, this is actually what I'm here for. Okay, all right, okay. Tell me about the melting wires in the DARPA chain. Right. These things never went on public roads, by the way.

These were experimental vehicles that we tested on various tracks for this robot race. But it was... I'll just call it. I mean, this was a seminal moment in... It's a huge moment. In the history of autonomous vehicles, you know, I don't want you to undersell your... Yeah, our audience is deeply aware of the DARPA. Okay, okay.

Participation today. It's just a robot race, you know. It was just a silly race that we did. No, we... I mean, launched a billion-dollar companies. It did. It showed the world what was possible, right? I mean, in 18 months, we went from a concept on a napkin to a vehicle that was able to navigate city-like streets, you know, have the right behaviors to handle uncontrolled intersections.

And, I mean, it was, we all, I think, surprised ourselves as to how well it worked. We had no idea what was going to happen. It was the first time we'd really had an opportunity for everybody to go build a robot to the same general set of rules of the road, throw them on a track and see what happens. And it was actually quite remarkable how well it worked. In terms of the hardware suite that you guys and most your competitors are using these days, you've got camera, you've got radar, LiDAR. Those seem to be sort of the Holy Trinity. Correct.

Is there sensing technologies out there that you guys are thinking about or looking at with some interest that maybe aren't part of that package so far, that maybe down the road there's a possibility of something else coming in? Because if you guys are talking about safety and redundancy, I'm sure you're trying to look for every possible thing that's out there that could help with that. Yeah, and the good news is that between those three, I think we have a pretty good, it really is kind of the Holy Trinity today. And the reason is because we see line of sight to build all three cost effectively at volume with all the functional safety requirements met. And that's a good thing, right?

Now, there are other modes out there. There's certainly thermal imaging cameras. Those are a little bit challenged because they're fairly high price today. Now, there are vendors out there that are looking at bringing that price down. They're looking at automotive qualifying it, but those are going to be on timelines that will not show up in the next couple of years. Those cameras are interesting. They provide interesting signal. In the past, when we've tested those types of sensors, they tend to, if you're in a really hot climate,

or if you're in a climate where the temperature, excuse me, changes over a very large range. it becomes a pretty interesting calibration problem so that the image doesn't look just fundamentally different day, day, night, which affects computer vision and all the other things that happen downstream of that to interpret that data. So it's not without its share of challenges, but I do think that's an interesting sensor modality, for sure. How limited are you by this sort of vendor ecosystem? You're saying it's a tightly coupled system.

How much is it we're like, actually we should invent thermal cameras, or this is what the industry is, we see the timelines we're going to invent the software suite that does sensor fusion of the other stuff. Yeah, so the timeline to build any new sensor from scratch is years, and that's just what it is, especially if you want it to be a reliable component that meets typical automotive standards. So that investment can't be taken lightly, and you want to know you're building it, that you're building something that will see the light of day, right? And because of all those factors, then, a lot of the tier ones are typical manufacturers that build this stuff, they don't take risks, they don't take huge risks. And so a lot of times what

you'll see is that we're hardening or making tweaks to things that came out of the driver assistance world. And in our case, we've done that, but we've also had to invest in, you know, from scratch development because the performance just isn't there. And this gets back to your earlier point around the driver assistance systems and how that's still equally as important as anything we're doing for the long run with autonomous vehicles. And I totally agree. the Tier 1 supply base still has plenty of work to do to get the sensors and computing systems where they need to be for driver assistance. And so a lot of the resources are on that, let alone being able to make that same resource available to work on longer range things for autonomous vehicles. So there's definitely a little bit of tension there.

Because obviously the goal is to get the cost down to a point where it's not prohibitive. And cost only comes down if you get the volumes. up. Right. So that gets us back to sort of like it's ready when it's ready. Obviously, it still needs to be ready at a point when you can start to scale and bring that cost down. When this unannounced Ford vehicle comes down.

Which is when? I'm going to keep trying. The thing about me is that it's never subtle. No, I like it. I appreciate that. It's right through the front door. Yes, yes. There's well Ford said 2021, right? See, that is subtle. Ford said November 4th,

2021. Is that correct? I'm remembering things. Nothing's changed. Okay. Actually, let's talk about Ford for a second. Why, you're looking at your PR guy? When is the hook coming? I just wanted Spide. I, I, I, he, he, he told, and I'm not going to go there. All right. Great. So now we can talk about your relationship with Ford. This is going to go perfect. You've got two big car. VW is the other one. Yes. Those are big companies to sell a lot of cars with ice engines and steering wheels. You're describing a vision that fundamentally disrupts their business. You're describing a vision that fundamentally says maybe people in city should not purchase a car. They should operate in a ride-sharing fleet with better congestion and environmental management.

What is that like in your conversations with Ford and VW? It's actually really simple. They don't sell a lot of cars in the urban cores because people are disincentivized to purchase a car. They don't want to pay the taxes, the maintenance, the insurance, the, parking. How much does a lease go for in Manhattan right now? Parking space for a car. Yeah, for a parking space. I saw the sign. I mean, I'll tell you how much I pay. I paid $2.75 a month. It's a lot of money. It must be... I don't live in Manhattan, though. Okay, okay. So in Manhattan, the sign I passed the other day was, it was something like

$5.99. And then there was an additional $180 tax on top of there. These numbers are not exact, but like, okay, that's a pretty big signal from the city that says, Don't bring your car here. It's right. And so because they don't sell a lot of cars, they realize that they need to, that the business needs to evolve. And it's not like personal car ownership's going away.

Of course not. People buy cars all over the place. It's just not with this particular demographic in cities because of the challenges we've talked about. So that means that the shift towards shared mobility services is happening. And how do they become part of that? And how do we make the economics of those shared services viable? The key is you need a few things.

You need the right vehicle design where we can give great experiences and move people around efficiently. You need autonomy so that it doesn't need parking spaces. It can just keep moving to go fulfill the demand in the area. And to do the right thing for the planet and everything else that comes with it, we need to make the switch and move to electric as well. So it's those confluence of things that are coming together that the automakers are saying, hey, we need to solve this problem. If we want to go after this market, we need to have.

those three kind of pillars as part of our business. What do you learn from like the riot hailing companies that have tried this? They've tried the sharing Uber pool, Lyft carpool. And it hasn't been as much of a success for those companies as they originally anticipated. It's just a sort of fact of human nature. People don't like to share things, especially American consumers, I would say, are probably less likely to want to share. Even when like the price point is at a highly competitively competitively.

credit rate. I mean, Uber was just dumping money into Uber pool trying to subsidize the heck out of it. And while it's still around, it's just not, it's clearly not as successful as they thought it was going to be. So how are our autonomous vehicles going to sort of solve that problem? So I'll try to tease this apart the best I can. There's a couple of things that capital gets sunk into in a ride-hale company today. And of course, I don't run a ride-hale business, right? So this caveat all this. But this is, I think, what we understand about these businesses is that First off, you don't look panicked. I can understand that.

That's how we know that you don't run a ride-haling company. You're not sweating profusely. Thank you for saying that. I'm not. I'm not paying. So the capital goes into a couple of places. Number one, in order to service demand, you need a supply of drivers. And there's a flywheel effect that has to happen here, right?

In order to drive and kickstart that flywheel, you need to provide subsidies to compel drivers to want to work for you to get on the roads. service that demand. So the more you have to subsidize those drivers, the more money you're spending on the driver. And that's what that's injuring the economics, right? And so if you don't need to pay a driver, then of course you are able to offer it with a more affordable price point, right? In theory. Now, this, this makes a lot of assumptions about the operation that you also don't have, you know, 10,000 remote operators driving the vehicle remotely. Right. Yeah. So, So it needs to be the right autonomous vehicle approach.

The second thing that Capital goes into is that these companies are, in addition to sort of driving the supply of drivers, they also have their autonomy efforts that they're piling money into. And this is where I think we're going to see some consolidation because that R&D spend can only happen. So can only continue for so long. because this in and of itself, building an autonomy engine, if you will, it costs so much money. It makes sense for it to be a platform. It does not make sense for just one party to have access to it. That's my belief. So you don't think people will want to personally own self-driving cars, or that will be a business model in the future?

I think it will be a business model once an autonomy-enabled shared fleet is successful. because you want the volumes, the volumes need to be up in a territory where the component cost comes down that an average person could afford it on their vehicle, on a personally owned vehicle. The third piece that I think is important to recognize is that these ride share companies are continually trying to expand. So if you go back to the first one on the flywheel effect, those subsidies get amplified if you're trying to take on, okay, now I'm going to go to India, now I'm going to go to Brazil, now I'm going to go to China. and you wage war potentially with your competitors, and that that continues to jack up those subsidies, right? And I think some of the companies have gone on record to say, look, if we stopped trying to expand and we just focused exclusively on profitability

in our most profitable regions, we could become a profitable company. So there's also this sort of delicate balance between growing and sort of growing the pie, but on the other hand, focusing the business in order to drive profit for shareholders. There's a lot of skepticism out there about self-driving cars. And it's, you know, I know that you're trying to do what you can. You have your own podcast. You're out there talking about things. And you talked a lot about education and wanting to make sure people are aware of this technology.

This is not a technology that, you know, the vast majority people have come into contact with. And yet it seems like a lot of people have already formed an opinion about it. And, you know, you can say that some of that had to do with Uber and the crash that occurred in Arizona a couple years ago. you can say some of that has to do with Tesla and the way that Elon Musk talks about these things and tends to sort of muddy the waters when it comes to talking about autonomy. How do you guys operate in a space where you have high levels of public skepticism and other players in the space sort of doing their own thing that doesn't necessarily maybe help alleviate people's skepticism about it?

So I don't know that there's broad skepticism. There's definitely pockets of it. I hear you. Look, I was in Miami a couple of weeks ago talking to a number of people in the community, right? And what I found was there is actually a substantial amount of optimism there. They realized that it's going to be very difficult to get new forms of public transit funded. They realized that there is a congestion issue and many of them don't want to drive either. And so it's one of those things. where I just think that if you pick the right city and if you explain it, and that is why we have our no parking podcast. Wow. The whole concept, thank you. The whole concept around it is to raise the education level in the cities that we're operating in, testing in, and give people an opportunity to just understand better what we're doing

and why we're doing it. And I find when we explain it and we explain it in a fairly accessible way and we talk about some of the advantages, people start to buy in. And I think any company that's in the AV business is also in the trust business, and they need to do their share of kind of bringing the community along. All right. So I ask every CEO, well, I'm doing it this year. I haven't done it historically, but I've started. Every CEO comes on the show.

I ask the same question, mostly because I want to know the answer for myself. When do you work? Because you've got to come on podcasts with Yahoo's like us. You got to do your own podcast. You apparently have to sweet talk mayors around the country into letting your cars on the road. you got to probably go to meetings. When do you sit down and look at stuff and say, okay, I'm making decisions or I'm producing something as an actual individual producer

as opposed to a manager? Yeah, my own personal tactic is I try to clump the travel together. So I'll put myself through a week and a half March to just go everywhere I need to go and do the outreach. It kind of gets me in a groove. It gets me out of the office. I enjoy it. But I do spend a significant percentage of my time at our offices engaging with people

and looking at the strategy and running the company. You just have to set really tight boundaries around how much time you spend and allocate to these things. You have to be disciplined in terms of how you set your time and be purposeful about it. What does that discipline look like for you? Do you have like calendar blocks or like leave me alone? People have given me all kinds of wild answers to this. I'm curious for everybody.

You should do a best of where you compile all the answers. Yeah, this is my pop business book in the making. It's good, it's good. Who move my cheese? Eli Patel. Dude, where is your cheese? That's my question. Yeah, I like it.

No, the range of answers to this question to me is fascinating. You know, I tend to look at it as a, I look at my schedule kind of like a map. So if you go to Google Earth or your favorite mapping application, you can zoom way out and see sort of the rough outline of things. I manage the rough outline. We're going to go to these offices. We're going to look at these sites. We're going to, this is where the travel will be.

and this is where I want to do one-on-ones for certain, to stay in touch with my leadership team. This is where we're going to have certain meetings. And then I let the team figure out how to fill in those blanks, right? When I go to Washington, D.C., I don't know all the people I need to meet with, but I know I need to be there. So we carve out the time. I set the high resolution, and I let the leadership team figure out kind of the microscopic details, and that tends to work really well. That's us, the microscopic details.

We are 100% a microscopic detail. You are one 60 minute block of many. Thank you. But a very important one. I've always wanted to be valued in exactly this way. This has been super fun, actually. There you go. See, we did it.

So my question is about context shifts, right? It's very hard, for me anyway, to be like, okay, I'm going to sit down and write an email now. Yeah. Or like, produce. And then I'm going to go to six hours of meetings and come out. This is something I've been particularly good at most of my life. I have a little bit. I can't do any one thing for.

too long. I love to do lots of things and it just keeps it different for me. Monotony is my enemy. So I love it personally. But my brain's wired that way and I know there's others that they're much better at going into the deep multi-hour thought session to come up with something that I'd never come up with. So we each play to our strengths, right? All right, man, tell the people, you already plug your podcast a little bit. Tell them about it again real fast. No parking podcast. It is hosted by myself and Alex Roy. You can find us on No Parking Pod on Twitter. You can go to No Parkingpodcast.com.

Please check us out. We're trying to raise the awareness of autonomous vehicles to people who maybe don't follow the subject every day and want to listen to some entertaining material at the same time. And what's next for Argo? Which people will be on the lookout for? What's next is we're going to be moving to the next step in terms of prototypes with Ford. We're continuing to expand in cities. We continue to serve our testing across a larger and larger area in the cities that we're in.

and we're excited to see what happens over the next couple of years. And this unannounced prototype. Yeah, that's right. Which we've now announced like four or five times. We didn't denounce any details. They've said plenty of times they're going to build a purpose bill of you. I'm just trying to. I'm literally trying to get the hook.

No, it's good. You're doing good. All right, man. Well, thank you so much for coming on the Vergecast. It's a great conversation. Thank you. It's a great time. All right, my thanks to Brian Selowski for talking to me and Andrew on the Vergecast.

We'll be back later this week with the chat show, the interview show next Tuesday and on and on and on. tweet at me on Matt Reckless. I love hearing your feedback. We'll see you soon.