Zoox thinks self-driving cars will change everything — just not all at once

Zoox CTO Jesse Levinson joins the Source Code podcast to talk autonomous vehicles, cars that aren't cars and kind of look like toasters, and what it'll take to win the robotaxi game.

Zoox thinks self-driving cars will change everything — just not all at once
Photo: Zoox

Jesse Levinson spent six years building a vehicle — don't call it a car, he'll hate that — before he decided to show it to anybody. The co-founder and CTO of Zoox has been working on autonomous transport for more than 15 years, and since 2014 has been designing and building a new kind of robotaxi for a driverless world. In December 2020, Zoox finally revealed its first product, and Levinson said the company's getting closer to letting regular people get inside.

Zoox was acquired by Amazon in the summer of 2020, but Levinson said that hasn't changed the roadmap at all. He's not trying to obviate the car industry or make sure nobody ever gets their driver's license again, but rather to reinvent the way people move around in cities. That's a big enough — and potentially lucrative enough — problem to keep the company busy for a while. After that? Who knows. Maybe Zoox will start delivering packages, if it can figure out how to get them to your front door.

Levinson joined the Source Code podcast to talk about the state of Zoox and the autonomous vehicle industry as a whole. A few years ago, companies were making grand proclamations about the immediate future, but Levinson and Zoox are a bit more conservative in the short run. In the long run, though, they're pretty sure autonomy will change everything.

Subscribe to the show: Apple Podcasts | Spotify | Google Podcasts | RSS

You can listen to the entire conversation with Jesse Levinson on this episode of the Source Code podcast. The below excerpts have been edited for length and clarity.

One thing that's always interested me about Zoox is that I think you and Zoox bit off a bigger chunk of the self-driving market than most. It would have been one thing to just say, "We're going to take a thing that looks like a car, that people sit in like a car, and it'll drive itself." You shift one thing and see what happens. And then maybe 10 years from now, shift another thing. But you did a bunch of pieces of it all at once. At the very beginning of Zoox, do you sit down to think, "OK, which parts of this do we have to do in order to make a dent?" Squaring that middle ground seems challenging to me.

That's a really good question. I think it's important to ask, "What is the least that we need to do to achieve the desired outcome with the constraints we have?"

A lot of people looked at us, especially in the early years of Zoox, and were like, "OK, I get the autonomy thing. You put some sensors, use some computers, write a whole bunch of software. But why the hell are you making a car? Like what is wrong with you people?" I mean, we got that from investors, we got that from competitors, we got that from car companies. People fundamentally just did not get it.

And the irony is, actually, our view has always been that we're making the problem easier. If you only solve a certain piece of the puzzle, then as I was saying, you're relying on other companies or other people to solve the rest of it. And while you don't want to take on too much — you want to do the minimum amount — you also need to control your own destiny.

We had a pretty strong view, even back in 2014, that if we just waltz into a car company and say, "Hey, you know what you should do? You should make a robotaxi." They're going to be like, "What? We sell cars." And cars have all these trade-offs. It's just a very different use case, when you think about the redundancy we need and the safety and the shape of the vehicle and the amount of battery so you can drive all day and all night long. And the customer experience being designed for riders, not drivers. These are fundamental departures from the automobile.

And so in our view, if you really want to build a safety-critical product that has very tight integration of hardware and software and the service model, you kind of have to own that architecture and the integration of it.

Even in the midst of doing all sorts of new stuff, though, you've been really focused on the idea of doing a robotaxi. A car has always been this thing that has to solve a million different problems for a million different people, while all sort of looking and working the same way. And rather than say, "We want to build a car that is useful for all the things that car might do," you picked a much narrower slice of the industry and said, "How do we build the best version of this thing?" It's not designed to do all the things a car can do. It's designed to do one thing super well. And my guess would be that that has been a useful focusing tool over time.

Yeah, it's been more than useful. I think it's been existentially important. It's just an incredibly hard engineering problem to get this thing all put together and demonstrably safer than humans. And the fact that we aren't trying to solve all things is crucial to that progress.

Right now there are 2.1 cars per family in the U.S. — we're not trying to solve that, in the sense that we're going to make it go to zero. Because then we'd have to do a million different things. And that would require many different vehicles. And hopefully, someday we have a dozen different kinds of vehicles, but that that day is fairly far in the future.

What we realized is that this one initial product that we're building is all by itself transformative for cities. And it could be a multi-hundred billion-dollar product. That's very unusual, right? There are almost no examples of single products that by themselves, with one model, could be hundreds of billions of dollars of value. So that helps us stay really focused, because we're not like, "Well, in order to be successful, we need these other 12 vehicles tomorrow."

So as you sort of noticed, we've kind of been just doing this one thing since 2014, when we started the company. And I think that's one of the reasons why despite having fewer resources than some of our competitors, in some ways we have been able to make more progress, because we've been consistently doing that one thing for almost seven years now. And I think we're maybe the only company that's been that consistent in our space.

That transition from where we are now is fascinating to me. And I want to know how you think about it, because it feels like we're in this position where even as people start to understand the idea that autonomous vehicles are safer by every metric than human driving, there's just some hang-up for a lot of people. There are lots of questions: If something goes wrong, who's at fault? What do I do? What's going on?

You can try and explain all of these things away by saying there are lots of problems when you're in a car with a person, too. But there's just something visceral and intuitive that feels better if I'm behind the wheel, even though it's demonstrably much less safe. Does it feel to you like that's going to be an easy hurdle to get over? Or is that a big one?

Yes and no. But as it turns out, we don't need everybody to want to try these vehicles on day one. I think we'll probably have the opposite problem, which is we're gonna have a ton of people really excited to try a robotaxi, and we can't just snap our fingers and make a million of them fall out of the sky. So there will be a ramp-up curve.

That said, it's absolutely incumbent on us to educate our customers and the rest of society, because it's not just the customers who are choosing to use this technology, it's the unsuspecting people that are sharing their roads with our vehicles. And if we can't quantitatively show that they're actually safer with our technology, even if they're not choosing to use it, we don't have the right to be on public roads.

As far as the way it feels for folks — and this is just our view — we think that the idea of getting in a car, sitting in the backseat and then there's just this disembodied wheel that's spinning itself with no driver? We think that's creepy.

Interesting! Because I was going to say, I would imagine just having a steering wheel, even if nothing is happening, is going to make it feel more familiar.

We'll see! It's an interesting psychology question. For me, if I'm in the backseat, and I see this wheel moving, it just reminds me that I don't have control over it. Like there's a missing driver. Whereas if the vehicle was never designed to have a steering wheel or a driver in the first place, it to me at least instills some sort of confidence: OK, this thing was designed not to have a wheel, and it knows what the heck it's doing.

How much does it seem like the industry has figured out about safety? We have great data, and the tests are getting better, but it feels like the regulations and the governing bodies and the certifications, at least as far as I can tell, are still kind of catching up to the New World Order. Where does it feel like we are in that process?

We've actually done pretty well as a country, in terms of being progressive with these with these new guidelines. So in 2016, NTSA came out with the world's first federal guidelines for fully autonomous vehicles. They said, change is coming, and on a federal level, we are going to consider that. And that was a pretty big deal. That was very forward-looking for them back in 2016. Many states have also paved the way for fully autonomous vehicles. And many more are following soon.

That's the good news. The bad news is we can't use the regulators as the scapegoat. We can't say, "Well, you know, we're all done and everything's great, but these stupid regulators are just taking their sweet time." No, they're actually not taking their sweet time. They're making this possible, and it's really on us in the industry to make the safety argument and say we are ready to put these things on public roads, because they're safer than a human. And that's really what we're doing.

That feels in a certain way like a very un-Silicon Valley approach. If you were to just take the traditional "move fast and break things" way of thinking about tech, all this stuff would be out on the road already. And it would be safer! There will be problems and things would happen — which is, to some extent, what has happened with Tesla and some others — but they're demonstrably better than having just a bunch of humans driving gas cars on the road. But it feels like you've been more careful than that. What's the thinking there?

I mean, the stakes are as high as they can be, right? Literally life and death. We also have an opportunity to create an incredibly valuable, important company for the long term here. That's an amazing responsibility and amazing opportunity. We're not going to squander that because we're excited to try something before it's ready to put on public roads. That would be a horrible miscalculation on every possible level, to prematurely release this technology. And I really hope that no company feels that they need to do that, which would not only destroy that company, but it would set the industry back.

There was the Uber fatality in Tempe, Arizona. And it came to light that there were a whole bunch of things that went terribly wrong at the same time, as is usually the case when something horrible happens. And it did sort of put a pause in the industry.

You know, it's not that this technology is going to be perfect. That's important to understand. We need to be much better than humans, but that's not going to be perfect. But if we can't very quantitatively, make the argument that we're significantly safer than humans, we can't be on public roads without a driver.

Where do you feel like the whole industry is right now? On the one hand, there's this narrative that the progress has slowed. Everybody was saying in 2015 that it was going to be 2020 when self-driving took over. And obviously that didn't happen. And so I think there's a little bit of a sense that we're at the bottom of the hype cycle now. But on the other hand, there's been all this consolidation, and there's money flowing through the industry and acquisitions are happening. What does it feel like on the inside?

I feel like we're a little bit past the bottom of that trough. The super hype was like, 2016, 2017. You had these companies putting out bullshit press releases, like, "Oh, we're gonna order 75,000 Volvos next year, and they're all going to drive themselves" or "We're gonna have 10,000 of these things driving in San Francisco in the next six months," and it's like, mmm, probably not.

And then people realize that like, OK, the "probably not" thing was a "definitely not." And then it's like, oh my God, this problem is impossible, humans are so great at driving, computers are never going to be able to do it, maybe it's going to be 30 years or 100 years. And I think that was an overcorrection to the over-optimism.

Now I think it's getting a little bit more grounded. You're starting to see that progress. And so I think you're going to keep seeing that. But again, people need to be measured. It's not that in six months, nobody's ever going to drive a car again. But it's also not going to take 50 years before you don't have to drive to work.

How has Amazon changed the trajectory of all that for Zoox? Obviously, you're part of a giant company that has quite a lot of money to spend on this stuff, but also has its own specific needs. It's not super complicated to imagine why Amazon might be interested in a large rectangular thing that could fit a lot of boxes inside.

You know, what's interesting is it's actually changed way less than I would have thought it would. When we first started talking to Amazon, we had the same thought you did: This is a company that moves a lot of stuff around, and that they'd probably be interested in moving around autonomously. It's a good idea! And it's not that they're not interested in that. They are interested in that. But to their immense credit, when we started talking to them last year, they actually got even more excited about what we're already doing, which is an entirely new market for them.

As a very large and successful company, they're always looking for ways to get into more markets and make the market more competitive. And so here was an opportunity, as a $1.5 trillion company, to do something that over time, could really make a difference.

It's not an overnight get-rich-quick scheme, we always remind people of that. But one of the things that's so great about Amazon is they're incredibly forward-looking. They're not afraid to think about what the world might look like in 10, 15 or 20 years. And very few companies can say that.

There are so many other great things about Amazon that eventually we can tap into, but they've been very clear with us that they bought Zoox to move people around cities. They don't want us to be distracted by all the other cool stuff we could do. Which is not to say we'll never do it, right? I mean, if you can move people, you can probably move other things, too. But what we're doing is already really hard.


Judge Zia Faruqui is trying to teach you crypto, one ‘SNL’ reference at a time

His decisions on major cryptocurrency cases have quoted "The Big Lebowski," "SNL," and "Dr. Strangelove." That’s because he wants you — yes, you — to read them.

The ways Zia Faruqui (right) has weighed on cases that have come before him can give lawyers clues as to what legal frameworks will pass muster.

Photo: Carolyn Van Houten/The Washington Post via Getty Images

“Cryptocurrency and related software analytics tools are ‘The wave of the future, Dude. One hundred percent electronic.’”

That’s not a quote from "The Big Lebowski" — at least, not directly. It’s a quote from a Washington, D.C., district court memorandum opinion on the role cryptocurrency analytics tools can play in government investigations. The author is Magistrate Judge Zia Faruqui.

Keep ReadingShow less
Veronica Irwin

Veronica Irwin (@vronirwin) is a San Francisco-based reporter at Protocol covering fintech. Previously she was at the San Francisco Examiner, covering tech from a hyper-local angle. Before that, her byline was featured in SF Weekly, The Nation, Techworker, Ms. Magazine and The Frisc.

The financial technology transformation is driving competition, creating consumer choice, and shaping the future of finance. Hear from seven fintech leaders who are reshaping the future of finance, and join the inaugural Financial Technology Association Fintech Summit to learn more.

Keep ReadingShow less
The Financial Technology Association (FTA) represents industry leaders shaping the future of finance. We champion the power of technology-centered financial services and advocate for the modernization of financial regulation to support inclusion and responsible innovation.

AWS CEO: The cloud isn’t just about technology

As AWS preps for its annual re:Invent conference, Adam Selipsky talks product strategy, support for hybrid environments, and the value of the cloud in uncertain economic times.

Photo: Noah Berger/Getty Images for Amazon Web Services

AWS is gearing up for re:Invent, its annual cloud computing conference where announcements this year are expected to focus on its end-to-end data strategy and delivering new industry-specific services.

It will be the second re:Invent with CEO Adam Selipsky as leader of the industry’s largest cloud provider after his return last year to AWS from data visualization company Tableau Software.

Keep ReadingShow less
Donna Goodison

Donna Goodison (@dgoodison) is Protocol's senior reporter focusing on enterprise infrastructure technology, from the 'Big 3' cloud computing providers to data centers. She previously covered the public cloud at CRN after 15 years as a business reporter for the Boston Herald. Based in Massachusetts, she also has worked as a Boston Globe freelancer, business reporter at the Boston Business Journal and real estate reporter at Banker & Tradesman after toiling at weekly newspapers.

Image: Protocol

We launched Protocol in February 2020 to cover the evolving power center of tech. It is with deep sadness that just under three years later, we are winding down the publication.

As of today, we will not publish any more stories. All of our newsletters, apart from our flagship, Source Code, will no longer be sent. Source Code will be published and sent for the next few weeks, but it will also close down in December.

Keep ReadingShow less
Bennett Richardson

Bennett Richardson ( @bennettrich) is the president of Protocol. Prior to joining Protocol in 2019, Bennett was executive director of global strategic partnerships at POLITICO, where he led strategic growth efforts including POLITICO's European expansion in Brussels and POLITICO's creative agency POLITICO Focus during his six years with the company. Prior to POLITICO, Bennett was co-founder and CMO of Hinge, the mobile dating company recently acquired by Match Group. Bennett began his career in digital and social brand marketing working with major brands across tech, energy, and health care at leading marketing and communications agencies including Edelman and GMMB. Bennett is originally from Portland, Maine, and received his bachelor's degree from Colgate University.


Why large enterprises struggle to find suitable platforms for MLops

As companies expand their use of AI beyond running just a few machine learning models, and as larger enterprises go from deploying hundreds of models to thousands and even millions of models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

As companies expand their use of AI beyond running just a few machine learning models, ML practitioners say that they have yet to find what they need from prepackaged MLops systems.

Photo: artpartner-images via Getty Images

On any given day, Lily AI runs hundreds of machine learning models using computer vision and natural language processing that are customized for its retail and ecommerce clients to make website product recommendations, forecast demand, and plan merchandising. But this spring when the company was in the market for a machine learning operations platform to manage its expanding model roster, it wasn’t easy to find a suitable off-the-shelf system that could handle such a large number of models in deployment while also meeting other criteria.

Some MLops platforms are not well-suited for maintaining even more than 10 machine learning models when it comes to keeping track of data, navigating their user interfaces, or reporting capabilities, Matthew Nokleby, machine learning manager for Lily AI’s product intelligence team, told Protocol earlier this year. “The duct tape starts to show,” he said.

Keep ReadingShow less
Kate Kaye

Kate Kaye is an award-winning multimedia reporter digging deep and telling print, digital and audio stories. She covers AI and data for Protocol. Her reporting on AI and tech ethics issues has been published in OneZero, Fast Company, MIT Technology Review, CityLab, Ad Age and Digiday and heard on NPR. Kate is the creator of RedTailMedia.org and is the author of "Campaign '08: A Turning Point for Digital Media," a book about how the 2008 presidential campaigns used digital media and data.

Latest Stories