Has the past year convinced you that your family doesn’t need multiple vehicles? Could you be persuaded to give up all of your cars? We may not be there yet as a society, but Jesse Levinson wants to get us closer to that reality.
A veteran in the world of autonomous vehicles, Levinson is the co-founder and CTO of Zoox, a Foster City, Calif., company developing its own electric, autonomous vehicle and robotaxi ride-hailing service. Levinson leads the company with CEO Aicha Evans. Zoox operates as an independent company inside Amazon after its acquisition last year.
On this episode of Day 2, GeekWire’s podcast about everything Amazon, Levinson talks about the acquisition, Zoox’s vision for environmental sustainability, and its larger goals in urban mobility.
Listen below, subscribe in any podcast app, and keep reading for an edited transcript.
The potential of autonomous, electric vehicles
Todd Bishop: Can you describe the vision for Zoox when it was founded in 2014, how the vision has evolved, and where you are today in terms of Zoox’s plans in electric autonomous driving?
Jesse Levinson: For sure. So we started the company almost seven years ago now, because we had this idea that the way people were getting around cities was just not great. And it was not great in a number of vectors. It wasn’t great for safety, because we lose 40,000 Americans every year in car crashes. It wasn’t great for people’s time, because people spend truly hundreds of millions of hours, actually billions, many billions of hours a year just driving and stuck in traffic, and it’d be great to be able to spend that time doing something more interesting. It wasn’t great for our cities. A third of traffic is just caused by people looking for parking. And a lot of cities are dedicated to parking lots and other things that it’d be really nice to get rid of over time. Certainly it wasn’t great for the environment.
And not only because gas powered cars are pretty obviously not a good thing, but actually even electric cars aren’t that great either, because you still have to make this thing. And they’re better, right? So given a choice between an electric car and a gas powered car, in almost every way that electric car wins out environmentally. But you’re still making a lot of stuff. And that’s not good. And so we said, if we could utilize our resources more efficiently, if we could have our vehicles driving customers around all day and all night long, and then spread out the economic and environmental impact of that vehicle over people using it all the time, that’s a major economic and environmental win. And then just from an efficiency perspective, people only use their cars 4% of the time. And 96% of the time they are literally just taking up space and depreciating. So that’s terrible, again, environmentally and economically.
Then you can look to maybe ride-hailing and say, “Well, what if we just use Uber and Lyft? And maybe the Uber and Lyft drivers can buy electric cars. Does that solve all of our problems?” And it doesn’t because then you have to pay somebody else to chauffer you around all the time. And that’s pretty extravagant, and very few people can afford to do that all the time. And so really the only solution that we saw was this idea of what’s now called robo-taxis, which are electric, autonomous vehicles that you can operate as a shared fleet and people can use when they want them. And when they’re not using them, somebody else is.
So that was our vision back in 2014. And critically, our view was that, to do that well, you needed a new vehicle platform. It wasn’t something that you could just retrofit onto existing cars by slapping some sensors on the roof and putting a computer in the trunk and writing some software.
So we’ve taken what most people perceive to be a harder route to solve the problem, with the view that, A, we actually think it’s going to be easier because when you’re building a safety-critical system you need to own the entire stack and be able to certify and validate it. And B, it’s going to result in a much better product that people are really going to want to use. And it’s going to be economically viable because we’ve made trade-offs that are designed not for the constraints of selling a car to a consumer, but for operating a fleet of robo-taxis in a city. So that’s how we kind of got started.
And the second part of the answer, which will be mercifully shorter than my first part, is actually very little has changed. We’re still absolutely 100% doing that. We’ve made a lot of progress. We’ve learned a lot over the years. The vehicle we shared with the world in December is really the culmination of that journey that we’d been on. And now we’re in the finishing touches to get it ready for public roads, and we make it incredibly safe and safer than humans. That’s also really hard, by the way. But that’s what we’re up to now, and we can’t wait for people to be able to use it.
TB: It’s funny, for some reason, whenever I think about the vehicle that you unveiled in December, as you said, I think of Dr. Dolittle, because it can go either way. It’s a Pushmi-Pullyu. But obviously that’s just one aspect of the design. Can you describe for folks, who may not have seen it or who may want to learn more about the underlying approach, what the vehicle is, how it works, and how it represents that larger vision that you just explained?
Jesse Levinson: Yeah, for sure. And for anybody who hasn’t seen it, I encourage you to take a look at zoox.com and see it for yourself, because it’s a little hard to describe it in a way that you’ll fully appreciate how cool it is. But essentially it’s a carriage, it’s a people-mover. It’s a compact, bi-directional, four seat electric autonomous vehicle. It’s a bit like a pod. Some people joke that it’s a little bit like a toaster. We think it looks better than a toaster, although nothing against toasters.
But fundamentally it was designed for a very specific purpose, which is moving people around cities, enabling autonomy. So you have, for example, sensors on the top four corners that can see 270 degrees per corner with camera, radar, and LIDAR, really important for solving the AI problem. And then internally you have a cabin. We call it a cabin because it’s really like your own private space to relax and enjoy yourself.
We like to say that Zoox is designed for riders and not drivers. First of all, we don’t have a driver, so that’s easy. And second of all, what we’re left with is the rider. And cars are so optimized for people to drive them, as they should be. I mean, that’s literally what you do in a car, is you drive the thing. But as a result, the passenger experience kind of suffers. And so with our vehicle we were freed from a lot of those constraints, and we were able to make a really beautiful experience for the interior passengers, whether you’re taking a trip by yourself, or whether you’re sharing a ride with others.
And that experience, at the end of the day, is what matters. I mean, the people using these things, they want to know that they’re safe, but they probably, 99% of them, don’t really care how it works. They just want to know that it’s a great thing to use and it gets them where they want to go. And on the way there they’re having a great time. And so that’s really what we were able to do with this vehicle.
What’s next for Zoox’s robo-taxi
TB: You announced and unveiled this vehicle in December. What are the next steps, and how much of the puzzle of Zoox does this one piece represent? Is this the whole of your vision, or do you plan to do other things in the future as well?
Jesse Levinson: Kind of both. I think it is the whole of our vision in the sense that really, again, why we started this company was to create a new way for people to move around cities. And this first product that we’re building is really the combination of many, many years of working on that vision. One of the things about Zoox that I think is really important is how focused we are. There are a lot of companies in the autonomous vehicle space trying, in my view, what might turn out to be too many things at the same time.
It’s easy to get tempted by, “Oh, we can move packages, we can do trucking, and we can move people, and we can do freeways. We can do this and that. And let’s have six different vehicles. We’ll have a big one and a small one and a tall one and a short one.” And all of these things are going to happen. And it’s good to have many companies working on parts of them.
No one company is going to own the entire AV landscape. That’s too big of a market for any one company to own. But, as Zoox, we have chosen to be very, very focused on this first product. Partly because it’s really hard. And we know that if we try to do too many things at once there’s a good chance we won’t pull any of them off. But second of all, and I think this isn’t necessarily appreciated enough, is the size of the opportunity, both socially and economically, with even just our first product, is truly massive.
You can come up with things that our first product doesn’t do, and that’s totally fair. It doesn’t do everything. It’s not going to take you Lake Tahoe from the Bay Area with all of your stuff. It’s not going to drive you to Los Angeles and back from San Jose. It’s not going to do everything for all people. It’s not going to be able to move 12 people at a time in one vehicle.
But it does a lot for a lot of people. And so there’s a multi-hundred-billion dollar economic opportunity with just this first product. And that’s why we’re so focused on it. Absolutely, over time, we will do other things. We will earn that right. But we’re not in a tremendous hurry to be doing seven other things, because this first one we’re working on is very worthy. And it’s, frankly, very hard. So yeah, we’re working on the safety of the thing. Really, really just handling the corner cases, making sure that across-the-board we are not just as safe as humans, but meaningfully safer than humans. And being able to quantify that. It’s not enough when you build a self-driving vehicle to just say, “Well, we did some experiments, and it seems pretty good. Let’s see what happens in the real world.” Right? That’s not appropriate. You have a lot of systems engineering and simulation and real-world testing to do to be able to quantify just how safe is this thing.
And again, you need to be able to say that it’s unambiguously and quantifiably significantly safer than humans before you unleash it. And that’s exactly what we’re working on now.
Backstory of Amazon acquisition
TB: You’ve had an eventful last year at the company, not the least of which was Zoox’s acquisition by Amazon. I recognize that many different factors go into a decision like that, not the least of them economics. But from a big-picture business and technology perspective, why did it make sense to become part of Amazon?
Jesse Levinson: It was a little bit unexpected for us, honestly, because we had been independent for quite a while, and we were excited about that. And the main reason we were excited about that was not so much because we’re like, “Oh, we’re so great, we don’t want to talk to anybody,” it was because we knew we were on a unique and important mission. And we had a feeling that most big companies maybe wouldn’t appreciate that. Or they would be like, “Well, here’s some technology. We can use it for this other thing.” And that might kind of be a bit of a waste. Not a complete waste, but not really living up to that full dream that we had when we started the company in 2014.
What was unique about Amazon was two things. The obvious one was they’re an incredibly successful company. They have wonderful financial and other resources to support something like this, and that was certainly not lost on us. But what was more pleasantly surprising, although in retrospect not really because you look at their history and how many really new technologies they’ve invented and created and really invested in over the long haul. They’re not a company that’s just looking for the next get-rich-quick scheme, was that they instantly saw that the vision was worthy. It wasn’t just a bunch of smart people with some cool technology that they could use to move packages around and save a few cents here and there. It was that, if you can move people around cities, you can enter this eventually multi-trillion dollar market, and really increase competition in a great way for consumers, and do something important, and hopefully meaningfully grow Amazon’s value over time.
And it’s hard, when you’re a $1.7 or $1.8 trillion company, or whatever they are today, you’re looking for those big opportunities. You’re also looking for small opportunities, to be fair. But there’s only so much that an extra five or $10 billion of value can move the needle on your market cap. And so you’re looking for those rare opportunities where you can create multi-hundred billion dollar businesses, and this is one of them. And they loved our approach, they loved the vehicle, they loved the way we were tackling this problem. And so to get to be a fully independent subsidiary of Amazon and continue to work on our full and unbridled mission with the resources and support of Amazon, was kind of a dream come true. It wasn’t exactly what we had anticipated six months earlier before the pandemic hit, but we were just really proud that they saw that opportunity and that they wanted to partner with us.
EVs, AVs, and environmental sustainability
TB: You also have some alignment, I know, on the issue of sustainability. Amazon had been criticized in the past for not moving faster to implement sustainable programs in terms of especially their logistics and operations programs. But in recent years they have announced the Climate Pledge to achieve net zero carbon emissions by 2040. Now many other companies, including Zoox , have joined that pledge. Where are you in terms of your work with Amazon on sustainability issues and Zoox’s big picture work on sustainability issues on its own?
Jesse Levinson: We are very aligned with Amazon’s view on this topic. And that’s, again, one of the things that really drew us to Amazon. I mean, it admittedly wasn’t our number one concern when we were finding a path forward was like, “Who’s going to be the most sustainable company?” Although absolutely we certainly didn’t want to be with a company that didn’t care about that. Our first concern at that point was like, “Okay, how do we secure our path forward and really get this out in the world?” But it’s actually no surprise that we’re so aligned with Amazon. Because again, sustainability is absolutely one of the core few reasons why we started this company in the first place. And so for them to recognize that and be excited about that was definitely part of the conversation even from the early days.
We are very much a part of the climate pledge. And that’s something that we’re excited about. It’s not like, “Well, Amazon has to do it, so I guess we probably have to do it too.” It’s like no, we’re already moving in that direction. I mean, we’ve been doing a lot over the years beyond just … I mean, we can all like pat ourselves on the back for the mission and how sustainable it is, that’s great too. But that’s not enough. Right? We have to be going out there and actually saying the right things and working with the right political and regulatory efforts and supporting the things that we think are important.
While it is absolutely the role of regulators and politicians to protect the environment, because there are externalities. And if you just say, “Hey, free market, do whatever you want,” and again, I’m a huge fan of the free market, but there are times where you have to put constraints in place to handle externalities. And I think the environment and protecting it is probably the number one example of that. And you can come up with plenty, we don’t need to diverge into politics. But protecting the environment, you can’t just say it’s a free-for-all and let’s see what happens. Because you can absolutely predict what will happen, and it’s not great. So to that extent we’re aligned with Amazon on some of this. But it’s one of these one of these opportunities where you don’t necessarily have to make some of the hard trade-offs that you otherwise have to make.
Sometimes you have to say, “Well, we’re going to do this thing, and it’s going to cost a lot of extra money, but it’s worth it because it’s going to be good for the environment.” And that’s great, by the way. I’m glad that companies are doing that sometimes voluntarily, and I’m also glad that the government makes sure that companies do that sometimes even if they might not want to. But what’s super-cool about Zoox is this is a win for the environment that’s also a win economically. It’s not one of these things where you have to kind of pick and choose, or you have to trade-off like, “Ah, we’re going to do this thing. And it’s such an expensive burden, but fine, we’ll do it.” The Zoox model is economically far superior to making and selling cars and having people own them. And yet it’s also so much better for the environment because you don’t make as much stuff. And so that’s for me what’s the most exciting aspect environmentally, is there’s not that unfortunate trade-off that there often is in business.
TB: I know that autonomous driving has been something you’ve been immersed in since your days at Princeton, and then Stanford. Do I have that right?
Jesse Levinson: That is right.
TB: And you were involved in the DARPA Urban Challenge, which is one of those legendary contests where engineers try to figure out the future of autonomous driving. But from a technical standpoint, how much of the environmental benefit do you expect to come from EV, electric vehicles, versus AV, autonomous vehicles? Are there efficiencies on the AV side that could potentially even trump the efficiencies you get from shifting from gas to electric? In other words, am I that bad of a driver? That’s really what I’m asking you here. Do we need a robot to drive us to save the planet?
Jesse Levinson: That’s a really great question. So the good or bad news, depending on how you look at it, is you’re not a horrible driver. And we’re not going to get like an order of magnitude to win by having the AI drive instead of the human, except we are, but in a different sort of a way. So let me explain that. It’s not that the AI algorithms are going to be so much more efficient at planning and breaking. And to be clear, they are going to be better at those things than you are. I mean, I’ve never seen you drive, but you’re probably not as good as our AI system in those areas. And that will give us some wins for sure. I mean, we’ll get 5% here, 10% there, 20% here. These are meaningful wins. But they’re maybe not completely transformative.
But what is transformative, beyond just being electric, and I touched on this earlier, but it’s such an important point for society, and I don’t it’s broadly appreciated yet, is it’s the fact that you have an AI driver in the first place that allows you to switch from the model of privately owned vehicles, whether they’re electric or not, to a shared fleet of electric vehicles. That is key for the environment. Because we we’ve looked into this a lot. And it turns out that making a car, and just the stuff you have to put into a car, including batteries by the way, right? I mean, batteries, aren’t great for the environment. There’s still a net win over the alternative, which is an internal combustion engine, burning gas. But everything else being equal, it’d be great to not build as many batteries. Right? There’s a lot of stuff in them.
And so it turns out that by removing the human driver and replacing it with AI, you can unlock that new business model of shared electric autonomous vehicles. And you can dramatically reduce the number of vehicles you need as a society. To put that in perspective, Americans own 2.1 cars per family right now. That’s an extraordinary amount of cars. We don’t want that many cars in our society for lots of reasons, including sustainability reasons. Now, we’re not coming out and saying, “Hey, nobody’s ever going to be allowed to drive a car or own a car.” As I touched on earlier, there are things that our first product will not do, and there are valid reasons to still want to own a car even if we were out in the market at scale. Maybe some number of decades from now that won’t be true, but it’s going to be true for a while. But I can tell you, that with this type of technology out there coming from Zoox, and I’m sure a couple other companies, we won’t need as many cars per family.
And we can start replacing privately owned cars with people using these shared electric autonomous fleets, and we’ll dramatically reduce the amount of stuff we have to build as a society. And that’s really where AI contributes to that major, major win.
TB: So as you say that, I think not only about my own experience driving, but also about my own experience over the past year in the pandemic when the shift to remote communication in the business world just totally changed the amount of travel all of us do. I ended up realizing, just as one example, and I think a lot of folks had the same experience, my family did not need two cars. How much has the past year reshaped the world’s awareness of that general issue that you just described?
Jesse Levinson: Maybe not as much as it should. I mean, you’ve had that insight, which is great, but you’re also thinking about these things more than most people are. But I think it’ll come. Right? Even though maybe not everybody has realized that, they have at least started noticing some of the ingredients that will feed into this epiphany for society. And absolutely, the advent of ride hailing from companies like Uber and Lyft in America, plus the pandemic and people realizing, “Hey, maybe there are some things I can do without having to physically go somewhere.” I think that is contributing to this realization over time. But it has to be said that there are lots of good reasons to go places. I don’t think we want to be a society that just only does things over Zoom. As wonderful a tool is that is, especially during a pandemic.
We do need this technology. We do want people to be able to get around. But as you say, it’s definitely not going to end up at 2.1 cars per family. And you’re just going to start to see that number drop and drop and drop. And I don’t know when it’s going to get to zero. But again, it doesn’t have to get to zero for it to be a huge win for society. If that can get to one, even, from 2.1, that’s a tremendous win, and that’s still a lot of people owning cars.
Package delivery and jobs
TB: I know that you’re focusing right now intently on human transportation, the robotaxi concept that you’ve rolled out and are moving forward with, even under Amazon, which I think should be noted, by the way, that’s really remarkable that Amazon was not doing an acqui-hire here or a technology acquisition. As you said earlier, you’re doing your own company inside Amazon. So I want to underscore that. But what are the potentials for package delivery, Amazon’s core business? What role will Zoox play in that? And what are the implications for jobs, because those drivers are not driving for recreation or personal enjoyment, they’re driving for income. How do you think about all those issues as they relate to what Zoox will do in the future?
Jesse Levinson: Those are good and fair questions, and we absolutely think about that. I think the first thing to put in perspective is that, as we’ve discussed, this is not an overnight, it’s not-a-get-rich quick scheme. It’s also not that society is going to look totally different in three months kind of a thing. Although we are making the leap to fully autonomous driving, the deployment of that technology will be done very much incrementally. And that’s not because we’re lazy or we’re not ambitious. It’s because it’s just one of the hardest technical problems of our generation. And it’s literally a safety critical device. And so that’s not something you can rush to market. That’ll quickly backfire. And so the reality is we as a society, and that includes people who make income from driving, have a long time to adapt to this change that’s coming. That doesn’t mean it won’t be difficult for some people, but if you do spend a lot of your time driving for income, you have many years before that’s going to be something you need to acutely worry about.
And maybe that’s something you might want to start thinking about now. But the demand for ride-hailing and for goods delivery is actually only increasing. So right now we have an acute need for more drivers actually. And yes, over the coming decades that balance will certainly shift. But again, usually when there’s a new technology it ends up creating more jobs than it displaces. And we certainly expect that to be the case here, in particular because we’re tremendously excited about the accessibility improvements that this technology will bring to underserved communities and people who maybe don’t currently have a safe way to even get to work. They might be in communities that are underserved by Uber and Lyft, both for economic reasons and maybe in some cases even for safety reasons. And both of those can be addressed by this technology over time. So we’re really excited about those benefits. As with all technology there are trade-offs and adjustments that have to be made over time. But again, we encourage folks to start thinking about that and realize that there’s plenty of time to prepare for this future.
TB: Many people in the audience are pretty familiar with the core tenets of Amazon. Some of them have worked there and have the 14 leadership principles memorized. I’m curious, is there anything that has surprised you about the company, positive or negative, that people who already know the company should know from your perspective?
Jesse Levinson: For sure. It’s been fascinating because we have kind of this unique experience where we are part of Amazon on some level, but we’re also not on some other level. So for example, we are separate in terms of our HR systems and a bunch of other facets and aspects. But we obviously have regular meetings with Amazon execs and get to learn from some of the fantastic lessons that they’ve learned and their culture over the years.
I’ll give you a couple examples. One of the things that we’re embracing to some extent, not because we’re being forced to, is the concept of the six pager, and they really like to write things down. And that’s really powerful, because sometimes you can just get lost in a slide presentation. And it’s easy to miss the point and the details. And so this idea that you actually sit down and you spend some time you write a three-pager or a six-pager, and then you give people time at the beginning of the meeting to read the damn thing and process it, and then have a conversation. That’s really powerful.
It’s not that we’re requiring people to do that. We don’t do it in all of our meetings. But we’ve started doing it in many of our meetings, even ones that aren’t with Amazon folks. And I’m a big fan of that. Another thing that I really like about Amazon and their leadership principles is this idea of really decomposing problems, looking at inputs and outputs. And asking yourself, if you say something’s hard, can you dig a layer deeper? And then what does that mean? And don’t just say, “Well, it’s unsolved, or, “Well, it’s hard.” Like quantify that, break it down. And then keep doing that recursively until everything makes much sense as possible. So it’s been great learning from some of the execs over there. And we’re certainly applying some of those lessons to the way we do things at Zoox. But we also appreciate getting to continue with our own culture and keep things still feeling like Zoox. So it’s been a really nice balance.
TB: Well, I will be looking forward to the day when I can just leave my car in the garage and walk out to the sidewalk and hail my Zoox.
Jesse Levinson: We are, too.
Subscribe to GeekWire’s Day 2 podcast wherever you listen.
