How Will Self-Driving Cars Change the World?

(phone ringing) Hello, sir. What's your name? - George. - George? Hi, how are you? So there's someone in the driver's seat. What is George gonna be doing? Nothing. Earlier this summer, I had my first ride in a self-driving car. It was at Waymo, the autonomous vehicle company that spun off from Google. Right beside me was Demetri Dolgov, Waymo's head of technology. So George is in the driver's seat, but he too is a near passenger in this driverless vehicle. That's right. Great. All right. I suppose I would say George take us away, but I suppose you're not driving. So let's go. I'll press the blue start right button. I press the button. Oh, and there's a beautiful chime sound. And again, we're already in the autonomous mode, so no inputs from me, no break. Wow, or staring. George, both hands are on his knees, and the wheel is moving itself. A decade ago, self-driving technology was a side show. Today, it's the new space race. GM, Ford, Uber, Tesla, they're all competing to develop the first fully self-driving vehicles. And nobody in the world is closer to that holy grail than Waymo. You see people walking around, we just passed a couple of folks on the sidewalk, and we are about to, I think, make a turn onto a bigger road. Yeah, I'm not experiencing heart palpitations, but this is definitely the most significant intersection that we've come up at. There's a left turn. Wow. This is weird. I mean, this is, it's strangely both normal and weird at the same time. It feels like magic and interesting, but then very quickly, it's actually pretty great how quickly people become comfortable with it. I have to tell you, I was so excited to drive around in an autonomous vehicle. When I got in, I was literally bouncing up and down, but the ride itself was kind of boring. The car drove itself perfectly, and that's just it. The car drove itself perfectly. The introduction of the automobile changed just about every category of modern life you can name. It injected speed into cities, opened the suburbs, changed pop culture, but of course, just as they were surprisingly good second order effects, they were bad ones. Dependence on oil polluted our skies and warped our foreign policy. The car was perhaps the most important technology in the 20th century, and everything about its relationship to the human is about to change. Autonomous vehicles are coming. The question is, what happens next? For the Atlantic, I'm Derek Thompson. This is Crazy Genius. The automobile has been pretty much the same ever since it was invented in the late 1880s by a carol bends. Lawrence D. Burns is a former executive at General Motors. Today, he's a consultant for several companies, including Waymo. He says, you might think a design that lasts 130 years has its merits, but actually the modern car doesn't make a whole lot of sense. We pay in the order of 30 to 35,000 dollars to buy a car, and we end up leaving it parked 90 to 95% of the time. That means three parking spaces for every one car in the United States. And when you include the time cost of your driving and your out-of-pocket cost, Americans are spending $4.5 trillion a year. $4.5 trillion a year. That's bigger than the US budget. And then a lot of people claim they love to drive, but in fact, driving is the distraction while someone's in a car going 70 miles an hour that weighs 3 to 4,000 pounds, unless they thought it was more important than the tax them to drive. Too big, too wasteful, barely used, and a waste of our attention. Ladies and gentlemen, I give you the automobile. It is really messed up, but yet that's the way it's been for over 130 years. In my mind, the biggest issue is safety. Today, more than 1.2 million people die due to car accidents worldwide. This is insane. That's Dimitri Doolgaugh again. At Waymo, he's trying to fix everything that's wrong with this inefficient death machine that rules our lives. When I went on that driverless ride with Dimitri a few weeks ago, we started in a parking lot of Google's office and Mountain View, California. I asked him to tell me how it all works. It looks like an ordinary minivan, except for a couple extra do-hikis. You see this big blob of sensors and a dome at the very top. This is what houses our main sensors, lighters, cameras, and radars, and as well as a set of microphones for detecting things like sirens from emergency vehicles. And there are other do-hikis along the back and front fender. What do you call this? It looks like a spinning device. Under the trunk of the car. It's a spinning device under the trunk of the car. It is a laser or a lighter. Basically, light detection and ranging device, it shoots out beams of light into the environment. And then it measures how long it takes those pulses of light to come back. So in a way, you could say that this is basically a bat mobile. But instead of echolocation, it's using lasers. You just summarize it perfectly. So that's really cool. You should just say that. Waymos are perfectly nice name, but it's a laser bat mobile is definitely cooler, I think. We could go with that if Johnny proves it. Meaning Johnny from Waymos Communications team. And well, he did not approve. But the car really is superhuman. Waymo isn't just trying to match our driving skills. It's aiming further. Our cars, they don't get distracted. They don't get tired. They don't fall asleep. They don't drive while drunk. They can see 360 degrees all around them, all at the same time. And all of their three sensor medalities, radars, lasers, and cameras. So would you say that this car can see as much as a human driver or more than a human driver? In some conditions that actually sees farther than a human driver would, at night, because lasers and radars emit their own light. We can actually sometimes see farther than you would be able. And similarly in bed weather, heavy rain or fog or snowstorm, they penetrate through those environmental conditions better than humanized. But of course, this isn't a public service. It's a business. And that's where things start to get even more interesting. Waymo sees four big ways of using this technology on our streets. Number one, ride hailing. Think Uber or Lyft, but without humans turning the wheels. Right now in Phoenix, select people can open the Waymo app on their phone and hail a self-driving car. Number two is logistics and deliveries. Long haul trucking, small good deliveries, just moving things around. Number three is public transit. And then the last one, number four, is personal car ownership where people would own a car that has self-driving technology in it. That self-driving tech in our taxis, in our trucks, and in our garages. We're not about building a car. We are about building a driver. It's a simple idea with huge implications. For more than a hundred years, we designed cars as driver-centric machines. It's why commercials talk about the thrill of acceleration, the feel of cornering. But self-driving cars will be designed around passengers, consumers. And that means they'll meet every need consumers have. Imagine a coffee shop on wheels that drives you to your doctor's office, or sleeping in a comfy hotel bed while it drives you to your in-laws for Thanksgiving. A self-driving power tool shed that pulls up to your curb whenever you need a drill to fix that broken door. If you think of driverless cars as just cars without drivers, you're not seeing the full picture. These are rooms with wheels. And that means they can be anything. Lauren Spurns again. Rather than selling a car through a dealership, you're going to sell subscriptions or rides and charge per experience or trip. In the future, according to Burns, we won't pay for machines that we drive. We'll subscribe to car services that drive us. After he left GM, Burns worked with the University of Michigan to see if this vision was even possible. He wanted to know. If you replaced all of the cars in a city like Ann Arbor, Michigan with electric autonomous vehicles that were shared rather than personally owned, how much would that cost? How many would you need? And how many privately owned cars could you replace on the road? We concluded that the 120,000 cars in Ann Arbor could be replaced with 18,000 shared electric autonomous mobility pods. This self-driving subscription service could replace 85% of privately owned cars in cities. They repeated the study in Austin, Texas, Columbus, Ohio, in Salt Lake City, Rochester, New York. Over and over again, if you had 15% of the cars in a city replaced by the shared fleet, you could serve the community, you could get to people quickly, even during rush hour. It's a thrilling vision of the future. Self-driving cars could be our multipurpose servants. They could be our chauffeurs and our subways, our hotels, movie theaters and our power tool sheds. And each car would be so tireless, so efficient that we wouldn't need a bunch of them in our garages and parking lots because wherever we were, whatever we needed, they'd all just be waiting there for us to summon them. I believe this future is inevitable. But not everybody is so optimistic. I see our future of autonomous vehicles as going either towards a heaven scenario or a hell scenario. A fork in the road will be right back. The developments that we are seeing in AI, this would be maybe the most fundamental thing ever really to happen. What's next for AI? And what does it mean for us? Values are getting exported through software. We need to have this conversation on how those values get exported. Join Atlantic CEO Nicholas Thompson for the most interesting thing in AI. Brought to you by Atlantic Rethink, the Atlantic's Creative Marketing Studio. Now on YouTube and wherever you get your podcasts. There is a point in the evolution of every technology where the question shifts from, "Can we do this? Should we do this?" For years people asked if Facebook could be a business or if Amazon could ever turn a profit. Today we ask, "Is Facebook an enemy of democracy? Is Amazon too big?" With self-driving cars, we're somewhere between questions 1 and 2. Yes, Waymo can drive a car, maybe even better than any human in the world. And yes, these vehicles can transform the urban economy. But should we want them to? To truly understand the risks of putting self-driving cars in our cities, you'd pretty much have to simulate an entire city. So, at the University of Michigan, they did. My name is Carrie Morton and I'm the Deputy Director of M City. And what is M City? Really, it looks like a fake city or some people say a movie set. And then we take those technologies out on the roads, onto the campus, here at the University of Michigan, and also the streets of Ann Arbor. And if I went there, would I see self-driving cars sort of snaking their way through fake city streets? Absolutely. Not only would you see cars driving themselves through M City, we have a low-speed driverless shuttle. We even have a fake pedestrian. His name is Sebastian. And he walks the streets of M City carefully. Wait, I'm sorry, a fake pedestrian. What is that? Students here at the University of Michigan created a mechanized pedestrian. Some people may use the term "target". We know these vehicles are not going to just be driving on the road, just amongst themselves. Real people will start to interact. We like to test with robot humans to make sure that they can safely navigate the roadway together. In the distant future, every car on the street might be autonomous. But in the transition, self-driving cars will share the road with us. Distracted, lazy, occasionally drunk humans. M City exists to make sure that transition isn't a total nightmare. As we drive down the road, we use a lot of non-verbal hand gestures and communication, hopefully nice ones. When there's a mixed environment where automated vehicles are in and amongst human driven vehicles, how do we replace that non-verbal communication? That's really important. If we again, think about these as part of a system. In an industry brimming with polyanas, Morton is a realist about how quickly self-driving cars will take over city streets. First, autonomous vehicles might be better drivers than humans, but they're very far from perfect. In March, an autonomous car from Uber killed a woman in Tente, Arizona as she crossed the street. Second, even if cars do drive perfectly, we may find ourselves with other challenges. Professional driving, whether it's trucks, taxis, or ubers, is the most common occupation in nearly every US state. Self-driving cars could replace these jobs and more. We even have talked about disruption, for example, in the hotel industry. Instead of flying somewhere, they may take an overnight trip in their car, which is outfitted with comfortable beds. So now you have two industries that are also going to be impacted by this technology. How do you see this affecting commuting, especially commuting from the suburbs? How much will that change? Even though Ann Arbor's lovely, lovely in the summer, maybe I'd rather live on Lake Michigan and commute two hours each way to work in my private pod. That is a likely possibility. It could play out that way. But add it up. The citywide fleets from Waymo, and Uber, and others. The self-driving trucks, the hotels, and coffee shops on wheels, and of course, the personal cars. Because let's face it, for a lot of Americans owning a car is practically a constitutional right. So add the suburban pods and the urban pods. It's so many damn cars. I see a future of autonomous vehicles as going either towards a heaven scenario or a hell scenario. Robin Chase co-founded Zipcar, a car sharing company that I myself find quite useful. But today, she's worried about a world where self-driving cars become too useful. I think about when we had smartphones at first-edited cameras, and you first thought, "What? Whatever. I don't need a camera, except for on birthdays and one of my vacation." I don't take a photograph every time. And think of how often we use those cameras today. So just picture the same thing is going to be true with your car. Give an example. How would a self-driving car create more driving for your family? My son, he was going someplace that was inconvenient by public transit. He was 17. I could say, "Hey, you, you're 17. Go take your bike or go take public transit." But no, I have a self-driving car. Hey, son, go use the car and then send it back. I tell the car to pick me up at 10, because I'll be finished with my breakfast, but I'm having a great conversation. So the car came at 10, but now it's just circling the block until I decide to come up, because I'm having my good conversation and whatever. Then I do go home, and I think I want to have pizza lunch, and so I send the car out to go get the pizza lunch, and it gets delivered back to me. So whatever errand you have in your mind, you'll now use this car to go do. And so that kind of comes up to this hellscape. A hellscape, aggravated by something economists call "induced demand." When you make something cheaper and easier, we use more of it. In fact, Uber and Lyft have already made traffic worse by encouraging people to hail cars where they might have once just walked or used public transit. One study found these companies have increased traffic more than 160% in downtown areas. Man, it's going to be 100% sedentary lifestyles that I sit at home, everything comes to me, and when I manage to go out, I'm just walking to the curb in front of my house. And remember, it's not just us ordinary folk who are making use of self-driving cars. Every company could see an advantage in putting their business on wheels. Amazon will do well to say, "Okay, I know the 1000 items that everyone buys, I'll put them in a car." And they're not going to be in some Amazon distribution warehouse outside the city. They're going to be roaming around in the cars. It's cheaper to have my inventory warehouse on public rights of way. And then, "Hey, Rob, I can give you 10 minute delivery, 15 minute delivery." This is the road to hell. If every family, taxi company, and retailer, once their own autonomous car, it will strangle our cities with traffic. Chase has a plan to fix it. If I think of the very first thing every city across America, and actually every city around the world needs to do now before self-driving cars gets in here, is we have to start putting a price on congestion. Congestion policies could mean charging a fee to cars that enter a city, or even capping the total number of cars that can pass through a dense area. That is, I do believe the single most important thing that we could be doing. I have to be honest, I loved talking to Dimitri Dolgov about his ambition to save 1.2 million lives from car accidents every year. And Lawrence Burns sold me on his vision. Self-driving cars do present us with a once-in-a-century opportunity to reimagine our economy. But I have no doubt that Robin Chase is right. Without a bit of thinking, self-driving cars will break our cities and worsen our national addiction to the automobile. This is a huge test for not only our technology companies, but also urban planners. It's one we face before. And we failed. I'm Eric Avila. I'm a 20th century US historian at UCLA. Avila Studies is another auto technology. One whose benefits and drawbacks might just predict the future of autonomous vehicles. But the tech key studies is so subtle, almost invisible, you probably forgot it had to be invented in the first place. The highway. You might think of the early highway system kind of the way that we thought about the internet 20 years ago. It was all about the idea of unfettered mobility, of freedom and migration and independence. What were the biggest upsides would you say of highways? highway stimulated the development of suburbs. People now had access, autonomous access in a way that they had not had before. Building a national highway system that centered around cities also meant a lot of destruction. Mass transit systems, street car systems were systematically dismantled. As public transit withered, highway congestion choked cities. Each white family fled for growing suburbs, taking tax revenue with them. The construction of urban freeways demolished poor and minority neighborhoods, like boilheights in Los Angeles. Today, when you drive through boilheights, you are literally surrounded by freeways everywhere. You are above freeways, you're below freeways, you're in the shadows of freeways, you can hear freeways, you can smell freeways. You're right there in your face. A new technology emerges in the US. It reshapes the suburbs, it connects the country, it grows the economy, but for millions of people, it turns the city into a hellscape. As it elevates the interests of the wealthy over the poor and the car over the human. Just about every hope and every fear for the future of autonomous vehicles is right here in the history of American highways. Automotive technology has already ruined our city's once in the last century. We can't let it happen again. I think the lesson is that we should not wholly rely upon the automobile as the sole and primary means of urban transportation. As long as the automobile is a part of our society, as long as it's enmeshed into the fabric of our lives, we're still going to grapple with the same problems that the automobile introduced half a century ago. Waymo says its goal is to improve driving to save human lives. That's noble, but it's not good enough. We should be aiming to replace cars too. For a century, we have designed our cities, our lives, our economy, around a machine that is inherently isolating and polluting, whose bulk creates congestion and whose presence saps energy away from what makes a city vibrant. If autonomous vehicles make us free, it will be because they freed us not only from the need to drive, but also from the constant need to be driven. It'll be because cities of the future were as bold and disruptive as the tech companies behind autonomy. They'll have to tax congestion, reinvest in public transit, yes, even if it's self-driving buses, use laws to encourage people to give up their private cars, and even ban ownership in dense cities. That's the only way to return cities to the people who actually live in them. Can you paint a picture of what the streets of an American city were like before the automobile? There were human interactions, human exchanges in a very promiscuous way. I don't mean that sexually, but I mean that socially. Cities sustained this promiscuous bustle of diverse people, of diverse pedestrians. In many ways, in many instances, the automobile destroyed that pedestrian promiscuity on the streets of American cities. As long as we're still talking about the automobile, I think there's reasons for both optimism and pessimism about the future of this new technology. Crazy Genius was produced by Patricia Yekub and Cashima Highlovic, with help from Asta, Chaturvedi and Kevin Townsend. For their expertise and help with this episode, special thanks to Jim McPherson, Olaf Sockers, Susan Shaheen and Mark Takahashi. David Herman is our engineer, break master cylinder composed our theme song and all the music in this episode. Catherine Wells is the executive producer of Atlantic Podcast. Special thanks to Matt Thompson. See you next week. [Music]