Are Human Drivers Finally Obsolete? | Freakonomics Radio

How a secret project at Google led to driverless cars on American roads. Freakonomics Radio shares a story from our friends at Search Engine. (Part one of a two-part series (https://freakonomics.com/is-this-the-end-of-driving/).) You can find the transcript and show notes for this episode on our website here: https://freakonomics.com/podcast/are-human-drivers-finally-obsolete/ FOLLOW FREAKONOMICS RADIO: YouTube: https://freak.ws/3yIl6dl Apple Podcasts: https://freak.ws/3yAvQh0 Spotify: https://freak.ws/3TsdCmV ABOUT FREAKONOMICS RADIO: Discover the hidden side of everything with host Stephen J. Dubner, co-author of the Freakonomics books. Each week, Dubner speaks with Nobel laureates and provocateurs, intellectuals and entrepreneurs, and various other underachievers to uncover things you always thought you knew (but didn’t) and things you never thought you wanted to know (but do) — from the economics of sleep to the future of education, from markets for marriage to the surprising utility of wolves. ABOUT THE FREAKONOMICS RADIO NETWORK: Freakonomics began as a book, which led to a blog, a documentary film, more books, a pair of pants, and in 2010, a podcast called Freakonomics Radio. Hosted by Stephen J. Dubner, it’s one of the most popular podcasts in the world, with a reputation for storytelling that is both rigorous and entertaining. Its archive of more than 500 episodes is available, for free, on any podcast app, and the show airs weekly on NPR stations. Freakonomics Radio is now the flagship show of the Freakonomics Radio Network, which includes the podcasts No Stupid Questions, People I (Mostly) Admire, and The Economics of Everyday Things. FREAKONOMICS RADIO NETWORK PODCASTS: Freakonomics Radio: https://freakonomics.com/series/freakonomics-radio/ No Stupid Questions: https://freakonomics.com/series/nsq/ People I (Mostly) Admire: https://freakonomics.com/series/people-i-mostly-admire/ The Economics of Everyday Things: https://freakonomics.com/series/everyday-things/ Special series: https://freakonomics.com/topics/ CONNECT WITH US: [email protected] https://twitter.com/freakonomics https://www.instagram.com/freakonomics/ https://www.facebook.com/freakonomics/ https://www.tiktok.com/@freakonomicsradio https://www.linkedin.com/company/freakonomics-media Subscribe to our newsletter: https://stephendubner.substack.com/subscribe

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:04Stephen J. Dubner?

  PJ, how've you been?

• 0:05Speaker 01

  I've been good. How've you been?

• 0:06Stephen J. Dubner?

  Yeah, I'm a little better now, having listened to your series. I love it.

• 0:10Speaker 01

  Oh, thank you.

• 0:12Stephen J. Dubner?

  Do you recognize that voice? It is PJ Vogt, host of the podcast Search Engine and friend of Freakonomics Radio. You may remember hearing him back in 2024 when we published a Search Engine episode called "The Fascinatingly Mundane Secrets of the World's Most Exclusive Nightclub" about Berghain in Berlin. That was a great story, and not too long ago PJ came to us with another one. It's a two-part series on driverless cars. This is a topic that we have touched on many times over the years at Freakonomics Radio, but PJ decided to go deep. The other day, I had a chance to ask him how he got interested in this.

• 0:50Speaker 01

  There's a whole lesson in this, but I'd gotten, and this is not the next sentence you're going to expect me to say, too into bench pressing.

• 0:59Stephen J. Dubner?

  Yeah, that's not where I thought you were going.

• 1:01Speaker 01

  And I injured myself. I had a hernia, and then I had to have a hernia repair.

• 1:06Stephen J. Dubner?

  I see.

• 1:06Speaker 01

  So there were like some minor complications. I was not moving easily. I was in a lot of pain. So I had kind of limited mobility, and I was visiting a friend in San Francisco, and I took a Waymo, and it was such an experience of the future that immediately becomes normal. First, the idea that I would press a button on my phone, a car would come out of nowhere, driven by nobody. I would get in, watch the steering wheel turn itself. I was trying to describe this to somebody recently. I was like, the first time it feels like the first time you're in an airplane, and by the third time, it feels like you're in an elevator. It was a moment where I thought, oh, a lot's about to change. And it was confusing to me that people were not talking about that more.

• 1:41Stephen J. Dubner?

  What should we expect to hear in the series? There are two parts. The first is really about the car, and then the second is really about the driver. Tell me who you think are some of the most compelling characters and why?

• 1:53Speaker 01

  So in the first part, there's this guy Sebastian Thrun. He's so good. He's this German-born roboticist AI expert who lost a friend as a teenager to a car accident. And he really thinks that his invention is not just going to make money for a tech company or be more convenient. He wants to reshape the modern world as it exists. And it's just the story of him and his team beginning to figure that out and having ideas that sounded crazy 20 years ago, and with every year towards the present have sounded more sane and at least plausible. And then in the second part, I find the Boston politicians to be very vivid talkers, very opinionated people.

• 2:35Stephen J. Dubner?

  Vivid is a very polite word.

• 2:38Speaker 01

  They're strongly opinionated. They sometimes commit gaffes. When you ask them about the gaffes, they are totally like, yep, I screwed that one up. The thing that I most enjoyed about this story, which is what I'm always looking for, is that particularly in the second half, every time I spoke to someone, as they were talking, I thought, everything they're saying makes sense. I totally get it. I would be nodding my head vigorously for the most part. And then I would go talk to the next person who saw things completely differently, and it would just spin my head the other way. And I would think, well, this makes sense, too. And it was about trying to really do what I think we're all going to have to do a lot of soon, which is weigh competing, not totally reconcilable interests, and really take them seriously. And that was us trying to do it for this one thing.

• 3:23Stephen J. Dubner?

  I'll be honest with you. I've been anti-human driver for about 50 years now.

• 3:28Speaker 01

  50 years?

• 3:29Stephen J. Dubner?

  Oh, yeah. I mean, have you ever seen a human drive a car, including yourself? We're not that good.

• 3:33Speaker 01

  No, I have no illusions about my driving skills. I'm not that good. I'm, I have a temper. I am distracted.

• 3:40Stephen J. Dubner?

  I rode in an autonomous test vehicle at Carnegie Mellon University. They had a test track in Pittsburgh on an old steel mill property. And after this one 20-minute or whatever ride, I said, give me the autonomous vehicles. It's so plainly better than I am as a driver, certainly. So I'm eager for it. And I appreciate your putting the pedal to the metal for autonomous.

• 4:05Speaker 01

  I hope it gives people a little context for this. All the questions people have, is it safe? What is it going to do? Like, we've answered as much as we can.

• 4:12Stephen J. Dubner?

  Today on Freakonomics Radio, we turn the mic over to PJ and our friends at the Surattention Podcast for the first of a two-part series on driverless cars. Listeners, start your engines.

• 4:51Speaker 01

  Before we start the story today, I want to ask you to imagine a different version of your life. You're you, but it's almost 200 years ago. And unfortunately, and I'm hypothetical, it's Monday morning. It's Monday morning and it's very early, pre-dawn. You wake up to this really hard rapping at your window. That's the knocker-upper, here to get you up for work. We're in the 1800s, before the invention of the adjustable alarm clock. The knocker-upper is a job. The knocker-upper walks the neighborhood with a long stick and taps it on the windows of people's houses early in the morning to wake them up for work. Who wakes up the knocker-upper for work? Nobody knows. But this is a job,

• 5:34Speaker 01

  a job that'll actually exist for another century. Outside, the gas street lamps are still burning. The lamplighter lit them the night before. He's supposed to come at dawn to extinguish them, but it's so early that he hasn't yet. Your lamplighter is one of those neighbors you have a deep fondness for, a fixture. Every day, you watch him make the rounds at dusk with his ladder and his light. You, yourself, are a driver. Professional driver 200 years ago is also a job. You're a person who sits on a coach and holds the reins of a horse. You take passengers where they want to go. You start your workday.

• 6:15Speaker 01

  Okay. Hypothetical over. Two of those jobs are obviously so long disappeared that most people don't know about them. The knocker-upper is your iPhone alarm. The lamplighter is the electric streetlight. The third one, driver, has persisted as a job for some, as a routine human task for nearly everyone else. This is a story about whether that's about to change. It's about how the word "driver", which right now makes me picture a human, could soon transform to refer to a machine, the same way the words dishwasher, printer, and computer all did. I've thought about this maybe too much in the year I've been working on this story. In conversations, constantly,

• 6:59Speaker 01

  I'd ask the humans I met the same question. Are you a good driver? Are you, do you consider yourself a good driver?

• 7:07Speaker 02

  I do, within limits. I think I'm a good driver because I understand the limitations of my driving.

• 7:17Speaker 01

  This is Alex Davies. He wrote an excellent book called Driven: The Race to Create the Autonomous Car. Alex, like me, thinks a lot about human driving, about his own personal limitations. What are the limitations?

• 7:29Speaker 02

  The limitations are that I can't always pay attention to everything, that I get tired. I've been trying really hard to be calmer on the road. My husband and I are expecting our first baby this fall.

• 7:44Speaker 01

  Congratulations.

• 7:45Speaker 02

  Thank you. And I thought that along with like, reading all the baby books, a good project to work on is just be calmer in the car.

• 7:53Speaker 01

  A very good resolution, because of course, for most of us, driving is the riskiest behavior we routinely engage in. In fact, even Alex, despite his good intentions, would actually get in a car accident just a few months after we first spoke. He was okay. It was the car that was totaled. Safety is the entire pitch for the driverless car, which is really a car driven by a computer. Driverless cars don't get drunk, tired, or distracted. They never text or feel road rage. And these driverless cars, they aren't the future. They're actually already here. But it's funny, if you just don't happen to live in a place that already has them, it's easy to not see how fast things are changing. Robo taxis, like Waymo, are operating

• 8:36Speaker 01

  in 10 American cities, providing millions of rides to Americans. In China, the rollout is happening even more widely. They're in twice as many cities. But here, if you live in a place like San Francisco or Austin, today, a driverless car is about as exotic as an Uber. A passenger in those cities opens up their phone and decides who should drive them: a human driver or a robot driver.

• 9:02Speaker 01

  How that happened is a story, a story we are living through right now, whose ending promises to totally reshape the places we live. And today, we're going to tell you how we got here, in chapters.

• 9:17Speaker 01

  Chapter one, Dreams Without Drivers.

• 9:21Speaker 01

  So it turns out this dream that inventors have had to replace the human driver with some kind of machine, that dream is about as old as the lamplighters.

• 9:30Speaker 02

  People have been thinking about a self-driving car for just about as long as there's been a human-driven car.

• 9:39Speaker 01

  Why?

• 9:40Speaker 02

  There's this funny thing you lose when you move from the horse to a human-driven car, which is that in a horse-drawn carriage, the horse is not just gonna run off a cliff if you let go of the reins. You lose sentience in your vehicle.

• 9:59Speaker 01

  When automobiles first arrived, these powerful and non-sentient cars, there's actually a passionate fight to keep them off the streets. It was the 1800s, and people feared these new things. The steam-powered vehicles thundering down the roads that soon evolved into gas-powered vehicles also thundering down the roads. The fear was partly about jobs. These vehicles were seen as a huge threat to a whole network of working-class jobs: horse breeders and horse farriers, horse feed suppliers, horse manure haulers, horse carriage manufacturers, not to mention the teamsters. Teamsters, today, the word makes me think of the Teamsters Union, but originally, the teamsters were the workers

• 10:42Speaker 01

  who drove teams of horses. Teamsters were like truckers before we had trucks. Cars seemed to imperil all these horse-related jobs. And even if you weren't worried about these workers, the cars were also less safe. Some anti-car activists battled to stop or slow the new technology, mainly with regulations. There were red flag laws, which said if you had an automobile, you had to hire a person to walk in front of it, waving a giant red flag to warn people. In Pennsylvania, a law was proposed requiring horseless carriage drivers who encountered livestock to stop, disassemble their car, and hide the parts behind the bushes. The governor vetoed it.

• 11:26Speaker 01

  But the thing about these crazy anti-car activists is that, directionally, they were right. Those cars did initially wipe out a lot of jobs, even if they created more. And cars were very unsafe. The cities that threw their doors open to cars without regulation were rewarded with astonishing death rates. Detroit let drivers pretty much run wild. In the early 1900s, deaths accumulated in a Detroit without driver's licenses, stoplights, or turn signals. Many of those deaths were children. It took decades for society to mostly learn to live with cars. The rest of the story is just the world you grew up in. We invented laws, licenses, driver's ed. We learned to better design

• 12:08Speaker 01

  roads. We invented the highway, the seatbelt, the airbag. All those things made driving less deadly, although the smartphone reversed some of that progress. Nationally today, deaths from cars are about as common in America as deaths from guns or opioids, about 1 in 100. It'll probably happen to someone you know in your life, maybe several someones. Whether or not you see that as an urgent problem to solve depends on you. But as long as there have been cars, there have been people who wanted to truly solve what's left of the safety problem, the best way we knew how. They wanted to make the car more like the horse it replaced, make the car more sentient.

• 12:49Speaker 02

  So that thought is there early, and like, early visions of it include, oh, we'll have radio-controlled cars, because they had radios at the time. There's a real effort at one point to build magnets under the road. And at each stage, what a self-driving car can be is dictated by the technology that's available at the time, for the most part. No one's thinking that much about a vehicle that thinks for itself. They're just thinking about a vehicle that the person in it doesn't have to drive.

• 13:27Speaker 01

  Many different attempts, many different failures. As many wonders as we invented, we could not approach nature's most majestic creation, a horse's brain. At least not until the turn of the millennium.

• 13:52Speaker 03

  Deep within the Department of Defense, there's a little-known military agency that has created some of the most innovative technology of the 20th century. This is the story of DARPA.

• 14:02Speaker 01

  Chapter two, DARPA's million-dollar prize.

• 14:06Speaker 03

  DARPA's current goal is to develop autonomous military vehicles, machines that can operate on their own without drivers.

• 14:13Speaker 04

  DARPA's always been intrigued with...

• 14:15Speaker 01

  This is from a documentary called "The Million Dollar Challenge." Honestly, less a doc, more an ad for DARPA, the Pentagon's research arm. DARPA's mission is to try to keep American technology one generation ahead of everybody else. It doesn't always work, but DARPA has invented or funded a lot: GPS and the M16, the early internet and the Predator drone. In 2002, DARPA decided to pursue the driverless car in a very unusual way.

• 14:43Speaker 02

  The director of DARPA, at the time a guy named Tony Tether, who had been a door-to-door salesman in his youth, definitely has that flair and that way of thinking, says, "Let's have a contest. Let's see who can put all of these ingredients that we've developed together into a proper self-driving car." His original idea is, "We'll drive them down the Las Vegas Strip." That's almost immediately next because it's insane.

• 15:12Speaker 01

  Oh, right. You would have to, like, literally gridlock a huge American city so people could put robot cars on it.

• 15:22Speaker 02

  Exactly. So he says, "Okay, do you know what? We'll do it in the desert. We'll do it in the desert, outside Las Vegas, and anyone who wants to can make a team, build a self-driving car, bring it to the desert, and we'll race them."

• 15:36Speaker 01

  The driver that DARPA wanted to replace was the American soldier. DARPA wanted a vehicle that could drive itself down roads that might be filled with hidden explosive devices. So, in this moment, at the tail end of the dot-com boom, DARPA's trying to inspire tech to build something besides another website. DARPA's Tony Tether announces that the prize for whoever can win its grand challenge will be $1 million.

• 16:00Speaker 02

  The rules were very open. There were little rules, like you couldn't have two vehicles communicating with one another. But you could build any kind of vehicle you wanted. Could have six wheels. It could be a truck. It could be a motorcycle. It could be a tricycle. It just couldn't attack other vehicles. That was ruled out early on.

• 16:18Speaker 01

  Oh, was that a concern that people would just like sort of battle bot the thing? Your autonomous vehicle would have like a little shredder that would take out somebody else's?

• 16:26Speaker 02

  Someone asked in the first Q and A at this, like they said, can we attack other vehicles? And they said, no. And it's funny you bring up battle bots because a lot of teams who entered this had battle bots history.

• 16:39Speaker 01

  Interesting.

• 16:40Speaker 02

  They were used to building robots for interesting purposes. And when they caught wind of this, they said, we can do this. We can scrap together some money and this will just be fun.

• 16:55Speaker 01

  I'm going to tell you what happened in this robot race in the desert. Not because I care so much about these early robot vehicles, but because I care a lot about the engineers who were making them. These would be the people who would later go on to lead development for the billion dollar companies creating today's driverless cars. And these people had very different views about how to get that technology ready. Different values when it came to things like the acceptability of risking human life. Abstract differences that would become very concrete later on, to the point where people would be charged with federal crimes.

• 17:32Speaker 01

  That's the future. But listening to this part of the story, what I listen for is how much of it can you detect already? How much are the differences already present? The first engineer I want you to pay attention to is a man named Chris Urmson. And way back in 2002, how did you end up being part of the DARPA Grand Challenge?

• 17:52Speaker 04

  Um, it sounded like fun.

• 17:57Speaker 01

  Chris, these days, the CEO of a large tech company. Back then, a PhD student at Carnegie Mellon University. When he first got recruited for the race, he was out in the field observing a robot as it crept across the Atacama Desert, training for its future deployment on the surface of Mars.

• 18:13Speaker 04

  My PhD advisor came down and was really excited about this DARPA Grand Challenge thing and the idea that you'd have a robot run across the desert at 50 miles an hour just sounded exciting, having spent. The last couple of weeks, walking behind a robot at very low speed.

• 18:34Speaker 01

  So Chris would join Carnegie Mellon's red team and help build a car called Sandstorm, a bright red Humvee with the top lopped off, a plethora of futuristic sensors mounted to it, like scanners a crackpot would use to search for aliens. You can see Chris back in that documentary. He explains to the filmmaker at the time that the hard part, of course, isn't the vehicle, it's the driver. How do you even begin to teach a computer to operate a Humvee at all?

• 18:58Speaker 05

  How does a computer make the steering wheel turn? How does a computer change the pressure on the brake and the throttle? Those are the issues that we're fighting through right now.

• 19:08Speaker 01

  Sandstorm represented the best entry from the contest's traditional academic crowd, but there's a different crowd there, too, represented best by a man named Anthony Levandowski. Can you tell me about Anthony Levandowski?

• 19:20Speaker 02

  Anthony Levandowski.

• 19:25Speaker 02

  Where to begin?

• 19:28Speaker 02

  So Anthony is like an entrepreneur. He's a really charming guy. He's six foot six. He's gangly, has all get out. He grew up mostly in Belgium because his mom was working for the EU. For high school, he moved to Marin to live with his dad. And he's a hustler.

• 19:53Speaker 06

  My name is Anthony Levandowski. I was a grad student at Berkeley. Instead of continuing on to finish my PhD, I decided it was much better to do the grand challenge.

• 20:04Speaker 01

  We asked Anthony for an interview. He didn't respond, but here he is in the footage from back then. Anthony did not have the engineering experience or resources of a team like Carnegie Mellon's red team, so he tried something very different, a vehicle that had almost no chance of winning the race, but which was also perfectly designed to stand out, to get him a lot of attention, maybe a job. The race's only self-driving motorcycle. It was named Ghost Rider, a stubby little thing covered in stickers with an antenna on the back and cameras on the front.

• 20:35Speaker 06

  There's a steering actuator on the top here, which allows us to modify the steering angle. So basically, if you're driving, you start to fall to the left, you steer left, that makes you turn to the left, and then you get centrifugal acceleration to put you back up to the right. And you're monitoring that in real time and making small adjustments and you stay balanced.

• 20:56Speaker 07

  The strobe light is on. The command from the tower is to move. Ladies and gentlemen, Sandstorm.

• 21:04Speaker 01

  The race happens on a Saturday in March of 2004.

• 21:08Speaker 07

  Autonomous vehicle traversing the desert with the goal of keeping our young military personnel out of harm's way.

• 21:22Speaker 07

  Booyah.

• 21:24Speaker 01

  What happens the first time they try to do this competition?

• 21:28Speaker 02

  The 2004 Grand Challenge is an utter hysterical disaster.

• 21:36Speaker 01

  Disaster number one, Ghost Rider, the motorcycle. Anthony Levandowski forgot to flip on the switch for the stabilization system. The bike immediately topples. Ghost Rider down.

• 21:48Speaker 07

  Anthony, good effort.

• 21:52Speaker 02

  And then every vehicle after it fails miserably. Like, one vehicle drives up onto a berm, flips off. One vehicle drives straight out, does an inexplicable U-turn, and just drives back to the starting line. And the rules are that, once your vehicle starts, you can't do anything.

• 22:11Speaker 01

  Even Sandstorm got stuck on a berm. Chris Urmson just standing there, unable to help his robot.

• 22:17Speaker 04

  Poor thing was trying to get going, but its wheels were just spinning on the gravel and tried so hard that it actually melted the rubber of the tires. And so there's just plumes of black smoke before they killed it.

• 22:30Speaker 01

  For the roboticists, this was obviously very disappointing. Chris Urmson compared it to an Olympic marathon, where the best runner only makes it two of the 26 miles. What this contest had done, though, was it had flushed all these inventors out. It had jump-started the scene that would develop this technology. One of the most important people there that day, actually just watching, was someone I haven't mentioned yet. A legendary roboticist named Sebastian Thrun.

• 22:57Speaker 02

  Sebastian Thrun, he was at the first Grand Challenge. He didn't bring a team, he wasn't participating. Darbo wanted to show off some other projects they'd been funding, including one of his robots, so he brings the robot. And so he's there, and he watches this disaster, and he thinks, "I can do better than this."

• 23:17Speaker 08

  I looked at the very first iteration of this grand challenge where I didn't participate. I was a spectator.

• 23:22Speaker 01

  This, of course, is Sebastian Thrun. He grew up in West Germany, moved to the US, taught at Carnegie Mellon before moving to Stanford. Watching that day, he saw this fundamental error he believed all the entrance had made.

• 23:35Speaker 08

  I saw that all the teams treated this like a hardware problem. They looked at this and say, we have to build bigger wheels and bigger chassis and so on. And I looked at this and said, well, wait a minute. The challenge really is to build a self-driving car that can drive through the desert. I can get a rental car that can do it just fine, provided there's a person inside. And the challenge is really to take the person out of the driver's seat and replace it by a computer. That is not a problem with bigger tires. That's actually really a software problem.

• 24:09Speaker 01

  Sebastian Thrun had a dual background: robotics and artificial intelligence, which probably explains his focus here on the robot driver's mind. He was thinking about something else, too. The military wanted this tech to replace a relatively small number of drivers in its war zones. But Sebastian was already imagining something bigger. What would happen to traffic deaths worldwide if one day everyone had access to a driverless car?

• 24:35Speaker 08

  I had experiences of losing people in my life to traffic accidents. And I felt we lost over a million people in the world to traffic accidents. Wouldn't it be amazing if DARPA invented something that would save a million lives a year?

• 24:47Speaker 09

  In October of 2005, 43 teams have brought their vehicles to compete in a unique event. A race driven not by testosterone, but computer code.

• 25:01Speaker 01

  Chapter 3. Machine Learning.

• 25:08Speaker 09

  The race course is a circular maze that zigzags for 132 miles.

• 25:12Speaker 01

  18 months later, for the second grand challenge, DARPA doubled the bounty, $2 million. This footage is from a PBS documentary called "The Great Robot Race", narrated, to my mild joy, by John Lithgow. Familiar faces have returned. Chris Aronson, back with the Carnegie Mellon team, this time with two vehicles, Highlander and Sandstorm. Anthony Levandowski, back with his motorcycle, which still doesn't work, he's knocked out in the qualifiers. And now, there's also Stanford's entrant. Compared to Sandstorm, the bulked-up Hummer, the car looks measly. A blue SUV, donated by Volkswagen. A baby-faced Thrun smiles next to his soccer mom-looking vehicle.

• 25:51Speaker 08

  The vehicle's name is Stanley, so Stanley is nothing else but Stanford. But it also gives the vehicle that personality. We think of the vehicle more and more as an intelligent decision-maker.

• 26:05Speaker 02

  And Thrun really brought more artificial intelligence, which at the time, we're talking 2005, was still rather primitive, especially compared to what we have today. But he could use it to teach his vehicle how to recognize the road and how to do it much faster. They found a dirt road out near Stanford, and they drive it down a dirt road and have the car's cameras record what they were seeing.

• 26:34Speaker 08

  The robot Stanley was able to train itself as it went. And the way it worked is, its eyes looked way ahead, and it could see stuff way at distance. When it drives over the stuff, it could tell, was it a good place to drive or not, because it could measure how slippery or how bumpy the road was. And it could then retroactively train itself and say, this green stuff over there, it's something good to drive on, AKA grass, and this brownish stuff, AKA mud, is not so good to drive.

• 27:03Speaker 01

  And so it was able to detect patterns and generalize from what it had learned?

• 27:09Speaker 08

  Yeah, absolutely. And it did this like 30 times a second, I mean, just like a person.

• 27:14Speaker 01

  The race kicks off with Stanley sandwiched between Carnegie Mellon's two behemoths.

• 27:19Speaker 09

  Highlander leads the pack, followed by Stanley and Sandstorm.

• 27:24Speaker 01

  What happens in the second race?

• 27:26Speaker 02

  The second race is as successful as the first race is disastrous.

• 27:33Speaker 01

  Nearly every entrant in the second race would go further than Sandstorm had in the first. Multiple vehicles would finish the course. The real question was who would do it fastest. And so at what point was it clear to you that you were going to win?

• 27:48Speaker 08

  Well, once we passed the front running team, we kind of saw the vehicle descend into what was the hardest part of the race course, a very, very treachery mountain pass. And we saw at a distance a dust cloud, we saw a helicopter, we saw little features that made us believe, wow, there's something happening that's magical. And this dust cloud then all of a sudden turned bluish because the car was blue and came closer. And then it came first to the finish line. It was unbelievably magical.

• 28:18Speaker 01

  At the end of the doc, over some criminally corny piano music, Sebastian Thrun gives his post-race interview. He's dressed a lot like a race car driver. Watching, you could forget he wasn't in the car.

• 28:28Speaker 08

  It was just amazing to see this community of people. That community succeeded today. Behind me, there are three robots that made it all the way through the desert, and all three of them did the unthinkable. It's such a fantastic success for this community. I think we all win.

• 28:48Speaker 01

  A made-for-TV kumbaya moment. Still years before the race to build driverless cars would enter its cutthroat phase.

• 28:59Speaker 01

  What would happen next is that a small band of lunatics would take driverless cars out of the desert, start secretly driving them on public roads in the state of California.

• 29:10Speaker 01

  They would do this at the behest of a man who'd been observing from the stands that day, disguised in a hat and sunglasses, who'd watched the challenge while his mind spun.

• 29:22Speaker 01

  That's after a short break.

• 29:29Stephen J. Dubner?

  I'm Stephen Dubner, and you are listening to a special episode of the podcast Search Engine here on Freakonomics Radio. We will be right back.

• 29:48Stephen J. Dubner?

  Hey there. It's Stephen Dubner. Today, we are running an episode from the Search Engine Podcast with host PJ Vogt.

• 29:55Speaker 01

  Chapter four, something actually useful for the world.

• 30:01Speaker 01

  The race in the desert had been designed as a spectacle, something flashy to draw out America's smartest roboticists. But it had drawn another person who'd come for his own reasons.

• 30:12Speaker 01

  Google's Larry Page arrived at the DARPA Grand Challenge in a baseball hat and sunglasses, a disguise. He found Sebastian Thrun and buttonholed him, asking him a million highly specific questions about things like the wavelength his LiDAR system used. But this meeting in the desert, this was not actually their first introduction.

• 30:31Speaker 08

  Well, the first time I met Larry was a bit earlier. He had built a small, little robot that acted as a telepresence for meetings, and he was trying to drive it around the Google offices instead of himself going to meeting with a robot. And he sent me a message and said, "I want to show you the robot I've built." And I, in a spirit of craziness, I sent a message back saying, "Larry, I'm so glad that Google lets you use 20% of the time to do something useful for the world."

• 31:00Speaker 08

  I couldn't. I either expected a rapid response or never hear from him again. It turns out I was lucky. He responded immediately. I took his robot, I fixed it next 24 hours, and he was very happy.

• 31:15Speaker 01

  Larry Page, it turned out, had actually been interested in autonomous vehicles since at least grad school. That's what he'd wanted to do his thesis on before being guided by some wise PhD advisor toward search engines instead. Now, as a spectator at DARPA's Second Grand Challenge, he could see real-world evidence that autonomous vehicles might actually be a thing. At first, Larry Page hires Sebastian Thrun along with fellow DARPA contestant Anthony Levandowski just to build what will become Google Street View. They'll actually modify the system that Stanley the Car's roof-mounted cameras had used to begin photographing American streets. But before long, Larry Page returns to Sebastian with his dream of a driverless car.

• 31:59Speaker 01

  And so how soon after arriving at Google does Project Chauffeur begin? Like Larry Page says to you, I have a mission? Like, how does this happen?

• 32:10Speaker 08

  This is an embarrassing moment for me. It's about two years later, 2009, where I sit in my cubicle and Larry Page comes by and says, Sebastian, I think you should build a self-driving car that can drive anywhere in the world. And my immediate reaction was, no, taking this technology we built for this empty desert and putting it in the middle of Market Street in San Francisco is going to kill somebody. And Larry would come back the next day with the same idea, and I would give him the same answer. And both of us got increasingly more frustrated, like, god damn it, it can't be done. And eventually he came and said, look, Sebastian, okay, I get it, you can't do it. I would explain to Eric Schmidt, the CEO at the time, and Sergey Brin, the microphoneer,

• 32:55Speaker 08

  why it can't be done. Can you give me the technical reason why it can't be done? And that's the moment of incredible pain, because I go home and I can't think of a technical reason why not. It was this kind of moment where I felt, look, I'm the world expert on self-driving cars, and I'm the person who denies that it can be done. Like, that taught me an incredibly important lesson about experts, that for the rest of my life, I decided experts are usually experts of the past, not the future. And if you ask an expert about innovation, something crazy new, they're the least likely person to say, yes, it can be done.

• 33:31Speaker 01

  So this is where the Google self-driving car project begins in 2009. It's led by Sebastian, joined by others from the DARPA challenges. The methodical Chris Armson was running most things day to day. Anthony Levandowski, the flashy motorcycle guy, would work on hardware. Dmitry Dolgov, another DARPA veteran, would be responsible for planning and optimization. It was a secret project. They'd report directly to Larry Page. A small enough team that there'd be no bureaucracy, few emails, fewer meetings. Just 11 engineers, who writer Alex Davies says, represented some of the best young talent in the country.

• 34:05Speaker 02

  And so Google builds this very quiet team. And it says to them, build us a self-driving car. And because that goal is super nebulous, they give them two challenges. They say, safely log 100,000 miles on public roads. But they also give them a challenge called the Larry 1K.

• 34:30Speaker 08

  So Larry and Sergey and I sat together and the two of them carved out 1,000 total miles of road surface in California.

• 34:38Speaker 02

  They open up Google Maps and they just click around and they look for 10 separate 100-mile routes that are really tricky.

• 34:48Speaker 08

  Absolutely everything, like the Bay Bridge and Lake Tahoe and Highway 1 to Los Angeles and Market Street and even Crooked Lombard Street.

• 34:57Speaker 02

  And they say to the team, you have to drive each of these 100-mile routes without one human takeover of the system, without one failure of the car.

• 35:07Speaker 01

  To get off to a running start, the team licenses the code from Stanford's DARPA Urban Challenge vehicle. Anthony Levandowski goes to a local Toyota dealership and buys eight Priuses, takes them back to Google and retrofits them to accept a computer as a driver. He hooks that computer driver electronically into the brakes, the gas, the steering. These Priuses get a radar system behind the bumper, cameras, a LiDAR system spinning 360 degrees on top. LiDAR, like radar, but it shoots lasers instead of radio waves. At first, the team gives each Prius a cool name, like Knight Rider.

• 35:44Speaker 10

  But I think we quickly realized that we're not going to be able to name all these vehicles as we scale up our fleet. And so we just started to number them like, you know, Prius 27.

• 35:52Speaker 01

  This is Don Burnett. He'd been a researcher working on autonomous submarines. He lost a friend in a car accident, separately got in a bad accident himself, and decided that he wanted to do work on self-driving cars. That's how he eventually ended up on the team in its early days.

• 36:08Speaker 10

  I was on the motion planning and behavior decision-making team, and my responsibility was to work on the nudging behavior.

• 36:16Speaker 01

  Nudging. When a big truck passes a human driver on the right, the driver will nudge a little to the left. For us, it's an instinct. Don's job was to teach a computer to nudge.

• 36:26Speaker 10

  You're trying to encode the behavior that you would use as a driver under kind of partially good perception. And it's a really tricky problem.

• 36:35Speaker 01

  A team of academic roboticists, some of whom had had friends die in cars, spending Google's money to see if they could make driving safer. It was a weird era.

• 36:49Speaker 01

  There's this big concert venue near Google's offices called the Shoreline Amphitheater. In 2009, you could have seen Sheryl Crow there, the killers, Phish. But the most interesting show that year was one almost nobody knew about. In the venue parking lot, on days when there was no concert, no tour buses around to see them, the Google team would run its first test runs of their driverless cars, essentially hiding in plain sight. A Prius, driving itself around the amphitheater parking lot with an attentive safety driver sitting behind the wheel, just in case. The team was making sure the basics functioned, that the sensors could really recognize another car, that the computer in the car was abiding by their orders. These were the baby steps that happened in this parking

• 37:33Speaker 01

  lot and at an empty airplane runway that was close to their offices. Spring 2009, the team tries actual, real road driving for the first time. Chris Ermson takes one of the Priuses out on the Central Expressway. Speed limit, 45 miles per hour. There are humans driving here. And immediately, outside the confines of the empty parking lot and empty airplane runway, here's what's clear. They had a real problem. The car was swerving wildly.

• 38:01Speaker 04

  It was weaving around like a drunken sailor. And we realized that the scale of the runway was such that you didn't notice the one or two-foot kind of oscillation it had in lateral control. And you put it on Central Expressway, and suddenly, you know, yep, turns out actually that's a problem.

• 38:22Speaker 01

  One more problem to fix. Listening to this story, it's funny because I can imagine it giving me a totally different feeling than it does. A tech company, with nobody's permission, was testing driverless cars on public roads in California. I don't know why that strikes me as being about invention, instead of just hubris and impunity. Maybe it's because I know that Google would be one of the few tech companies whose driverless cars would not cause any fatal accidents in testing. And that the team would just take more safety precautions than the other companies, who'd rush in later to catch up with them, once this was an arms race.

• 39:05Speaker 01

  The way these cars were designed, the safety driver sat behind the steering wheel, ready to take over. In the other seat was their partner, watching the monitor displaying a graphical interface, designed by Dmitry Dolgov. The people watching the screen would call out problems ahead, some discrepancy between what the sensors were seeing and what was actually in the road. This is what teaching a car to drive actually looked like. Two-person teams manning the cars, logging errors, going back to the office to troubleshoot, and then updating the code. I asked Don Burnett about this, Sarah. And while you're doing this, and then like you leave work and you get in your car that you drive as a human, did you find yourself thinking more carefully like, how do I know what I know when I'm driving? Like, you're trying to teach a machine by day. Did it

• 39:49Speaker 01

  affect how you thought about human driving by night?

• 39:52Speaker 10

  Almost obnoxiously so to any passengers in the car with me. I was obsessed with one big question, which is, why do humans drive the way they drive? And it turns out, there were no good answers. And I still think they're not great answers. And instead of actually answering that question, we've just turned to machine learning to infer the deep truths behind why humans do what they do. And so there's some basic principles that you can understand, like, we try to minimize lateral acceleration, meaning you don't want to be thrown to the outside of your car when you're making a turn. So you're going to slow down. But how much do you slow down? Right? And it turns out, that's contextual.

• 40:34Speaker 01

  Don gave me an example. So you're trying to figure out the right speed and angle for the car on one of those tight, curvy onramps onto the highway. You want it to feel comfortable for a passenger. Don says you can work out the math, the lateral acceleration is 2 meters per second squared. But the surprising thing is, that number only applies on the on-ramp.

• 40:56Speaker 10

  If I put you at a cul-de-sac in a neighborhood, and you were going to do a U-turn at the end of the cul-de-sac, even though the speed is significantly slower, if you did 2 meters per second squared of lateral acceleration around a cul-de-sac, you would tell your driver they were crazy. It would feel incredibly uncomfortable, like incredibly uncomfortable.

• 41:21Speaker 01

  You would feel like you were in Mario Kart.

• 41:23Speaker 10

  Yes, it would feel Mario Kart. And remember, this is a force. So it's a physical feeling on your body is exactly the same. But the contextual awareness of the situation of speeding up to get on the highway versus making a U-turn in a residential street tricks your brain into feeling Opposite about the situation. And so it turns out the limit for a cul-de-sac is around 0.75. It's almost three times less than you would be willing to tolerate as you accelerate onto a highway. And so there were things like that where you couldn't just say humans have specific physical restrictions, right? From a forces perspective, the context matters.

• 42:07Speaker 10

  And when the context matters, now all of a sudden, anything is game. So it, things like that is where I spent my time as a researcher, trying to figure out, okay, how are we going to make this comfortable for passengers?

• 42:19Speaker 01

  All these little problems to solve. But there was one gift, which was that the team at this point had an overarching goal uniting them. The DARPA challenge had told them, drive across this patch of desert. The Larry 1K challenge told them, drive these 10 routes without human intervention. The specificity of the mission meant they never had to squabble about why they were there. By 2010, just a year in, the team was really on a roll.

• 42:45Speaker 02

  They start knocking out routes.

• 42:48Speaker 10

  Each one of the routes was unique and distinct and different and had its own challenges.

• 42:53Speaker 02

  Down Route 1, Silicon Valley to Carmel.

• 42:57Speaker 10

  The bridges run. We had to go across all of the bridges in the Bay Area, starting in Mountain View, finishing crossing the Golden Gate Bridge.

• 43:05Speaker 02

  It's Chris Hermsen in the car. It's Anthony Levandowski in the car.

• 43:09Speaker 10

  I was in the car with Dimitri, Chris and Anthony. It was the four of us in the Prius.

• 43:14Speaker 02

  They're figuring out the technology much faster than they thought they could.

• 43:18Speaker 01

  The Larry 1K was set up like a video game, meaning they'd get to try the route over and over until they could complete it without a single human takeover. Then they'd move on to the next one.

• 43:28Speaker 04

  It was really a proof of concept exercise. Can you even make this happen once?

• 43:35Speaker 02

  When they fail a route, they know what the car can't handle. So they go back and say, you have to be better at doing X, Y, Z.

• 43:42Speaker 10

  And then we got back to the office. We regrouped. We went back out, I think at like 11:00 PM. And by 1:00 AM, we had completed the route.

• 43:52Speaker 02

  They buy a bottle of Corbella champagne. They all write their names on it.

• 43:57Speaker 01

  Corbeil, $13.99 a bottle, the champagne they have at Trader Joe's. They had one for every route they completed.

• 44:03Speaker 02

  And one by one, they pick off the Larry 1K routes. And they think this is going to take them about two years when they start out. And they do it in a little bit more than a year, nearly twice as fast as they had expected.

• 44:18Speaker 01

  By fall of 2010, they're done. Here's Chris Armson.

• 44:21Speaker 04

  And I think we had a big party up at Sebastian's house in the Les Lautaux Hills. So, you know, it was pretty spectacular, right?

• 44:28Speaker 02

  They throw each other in the pool. They celebrate and then they're not entirely sure what to do next.

• 44:36Speaker 04

  It was kind of, okay, and now what?

• 44:40Speaker 01

  The team had pulled off a kind of miracle in a year. A driverless car with human supervision, with lots of human coding, but still a driverless car successfully navigating some very tricky roads in California. They'd done this safely. They'd done it quickly. And now things would begin to wobble. Competition would arrive. The team itself would begin to schism. And one member, a person who believed the team was moving too slowly, would actually take matters into his own hands in a particularly extreme way. After the break, mutiny.

• 45:19Stephen J. Dubner?

  Hey there, Stephen Dubner. That again is PJ Vote and you are listening to an episode of the Search Engine Podcast here on Freakonomics Radio. We will be right back.

• 45:39Speaker 01

  Welcome back to the show.

• 45:44Speaker 01

  As early as 2010, Google's driverless car project had developed some very impressive self-driving technology. But what they were struggling to decide was this. What was the actual product they were developing here? Here's Sebastian Thrun.

• 45:58Speaker 08

  We had a lot of debates inside Google what their right business model was. At some point, we actually had a big debate. We should just buy Tesla. And Tesla was worth $2 billion at the time. I remember this.

• 46:11Speaker 08

  Maybe we should have in hindsight. But joking aside here, there was a debate whether this is more of an assistive technology or a disruptive replacement technology.

• 46:24Speaker 01

  Basically, should they follow the route that Tesla ultimately would? Design self-driving as a feature in your car, something that could take over sometimes, but still need human monitoring? Or was it better to wait until the car could fully drive itself? Thrun would eventually come around to this version of self-driving. Specifically, he'd come around to the idea of self-driving robo-taxis.

• 46:45Speaker 08

  A taxi service type system is way more capital efficient than ownership. An owned car is being used about 4% of the time and is parked 96% of the time. Imagine a city without parked cars, where every car is being utilized, call it 50% of the time, which means we have like only 10% number of cars needed that we need today when we own cars. That's going to happen, there's no, absolutely no question.

• 47:09Speaker 01

  What Sebastian is describing here so matter-of-factly is a fairly radical reimagination of American cities. The idea that robo-taxis would be so cheap and widely available that most people just wouldn't own cars, that we could put something else, anything else, in the places where we put most of our parking lots and parking spaces? That is a far-fetched idea, just given how much of American identity is tied into personal car ownership. A far-fetched idea, and for it to begin to happen, Google would have to bring a product to market. But the years passed, and they didn't. And some people who were there felt stuck. Don Burnett says he believes life at Google got dangerously cushy.

• 47:53Speaker 01

  The food was great, the money was too. These former academics making much more than they'd ever expected.

• 48:01Speaker 10

  There was a lack of urgency on the team to actually make something viable. We had a funding supply that effectively felt infinite, and maybe it was, maybe it wasn't, but it certainly felt infinite. And when you have infinite funding, you're not forced to make hard decisions, you're not forced to focus, you're not forced to look at the opportunity, the market, the customer, and be the best. It was more like, hey, let's take our time, let's make sure we do it right, which is, on its face... A good principle, but at the end of the day, I think the lack of urgency wasn't for everyone.

• 48:40Speaker 02

  And within the team, you get Team Chris and Team Anthony, and they start butting heads all the time.

• 48:47Speaker 01

  Chris and Anthony, meaning Chris Urmson, official head of the project, versus Anthony Levandowski, who I still think of as the motorcycle guy.

• 48:54Speaker 02

  The main difference in their approach is how quickly they want to move. Anthony is very okay with risk, we'll say.

• 49:05Speaker 02

  He gets one of these cars and he's driving it back and he lives in Berkeley, he works in Palo Alto. He's just using this car like on the Bay Bridge every day, probably outside the bounds of what the team actually wanted.

• 49:18Speaker 01

  Yeah.

• 49:18Speaker 02

  And he's not like necessarily logging data. He's just enjoying his self-driving car and taking it all over the place. Chris comes from an academic background. He's that Canadian, very nice, very careful, very risk averse.

• 49:33Speaker 01

  When I asked Chris Urmson about all this, his memory was slightly different. In his memory, Team Anthony was pretty much just Anthony. And Anthony, he said, was a move fast and break things kind of guy. Move fast and break things, a motto famously coined by Mark Zuckerberg. It defines a way of developing technology which once might have felt cute and revolutionary, but which today, at least to me, feels pretty irresponsible.

• 50:00Speaker 01

  Chris didn't think that philosophy was an option for their team. Even if their cars were statistically safer than human drivers, he knew that the first news story about a self-driving car in a fatal accident was going to be a huge deal. Anecdote was going to demolish data if they weren't extremely careful. By all accounts, Anthony Levandowski felt differently. But he actually wasn't the only one. Here's Don Burnett.

• 50:27Speaker 10

  There were some people on the team, very famously, including myself, that started to get the itch kind of towards the three to four year mark. The itch of like, okay, where is this going? Who is it for? How are they going to use it? Where are they going to use it? And I felt like the leadership didn't have great answers to that. There was no commercial race, right? We had no competition and there was no market for the product.

• 50:50Speaker 01

  But competition would soon arrive in the form of Uber.

• 50:58Speaker 10

  This was the oh shit moment for me. Uber announced their self-driving program. And I remember, like it was yesterday, waking up, reading the news, going to my desk in the morning and thinking, oh crap. These guys are going to eat our lunch.

• 51:15Speaker 01

  In 2013, then CEO of Uber, Travis Kalanick, had gotten a ride in one of Google's prototype driverless cars. Sitting in a taxi without a human driver, he'd understood that this could be the end of his company. And so Uber had plunged headlong into the driverless car race. The company hired nearly half of Carnegie Mellon's top robotics lab. And not long after, we also know through court records and emails, that Uber also began communicating with Anthony Levandowski, who in 2016 would leave Google, quitting just before he could be fired for recruiting team members away, including Don Burnett.

• 51:53Speaker 01

  Anthony would then start his own autonomous vehicle company. Uber would soon buy that company for almost $700 million, even though the company had no product and was only months old. Which raised a mystery. Why would Uber pay so much for a company whose only asset seemed to be its people?

• 52:12Speaker 02

  This is where Google goes into its computer security logs and realizes that not long before he left, Anthony Levandowski downloaded something like 14,000 technical files onto his computer and moved them onto an external disk.

• 52:28Speaker 01

  Obviously, you can't do that. I mean, I'm assuming obviously you can't do that.

• 52:32Speaker 02

  No, you definitely cannot do that.

• 52:39Speaker 02

  And this is the kind of thing that maybe if he had stayed there, this is the kind of thing Anthony would have done. And he would have been like, oh, it's just so I could have access to it somewhere else. And he probably would have gotten away with it. But when you then go and work for Uber and start running their direct competitor self-driving car program, that's when you get in trouble. And that's when what's technically called Waymo at this point, Google's program, sues Uber and puts Anthony at the center of an enormous legal battle between these tech giants.

• 53:16Speaker 11

  Secrets and subterfuge in Silicon Valley. A former Google engineer has been charged with stealing files from Alphabet's self-driving car project and taking them to Uber.

• 53:26Speaker 12

  Specifically, it involves a former lead engineer of Google's self-driving car unit, Anthony Levandowski. Now, he's accused of using his personal laptop and downloading more than 14,000 files.

• 53:39Speaker 01

  In 2016, Google had just spun its driverless car unit into a new entity, Waymo. Waymo sued Uber. Uber had to settle to the tune of $245 million. And in a separate criminal trial, Anthony Levandowski pled guilty to stealing trade secrets. Afterwards, Uber continues their driverless car program without him, continuing to pursue its move-fast, break-things strategy, which in 2018 leads to the death of a woman named Elaine Hertzberg.

• 54:07Speaker 13

  Uber is hitting the brakes on its self-driving cars after one of them hit and killed a woman in Arizona.

• 54:13Speaker 14

  The vehicle was in autonomous mode, but it did have a safety driver on board. But a police report later indicating the safety driver was streaming TV shows on her phone for three hours that night, including at the time of the crash.

• 54:28Speaker 01

  The way this story was reported, nearly everyone blamed the safety driver. She was on her phone. She was streaming an episode of The Voice.

• 54:35Speaker 14

  Tempe investigators saying had Vasquez been paying attention to the road, she could have stopped the car 42 feet before impact. The NTSB slamming Uber.

• 54:46Speaker 01

  There was some important additional context, which is that Uber's robot driver was also just much worse than Waymo's. A statistic I found jaw-dropping. At this point, Waymo's safety drivers were having to take over from the car once every 5,600 miles. Uber's safety drivers that year had to intervene more than once every 13 miles. Despite that, five months before the crash, over employee objections, Uber had cut its safety crews. Instead of two humans, they just used one.

• 55:18Speaker 01

  One safety driver overseeing a robot driver that was arguably not ready to be on public roads. In the last moments of Elaine Hertzberg's life, the robot spent an indefensible 5.6 seconds trying and failing to guess the shape in the road that was a human body pushing a bike. Over those 5.6 seconds, the robot kept reclassifying her. Was she an unknown object, a vehicle, a bicycle? During that time spent wondering, the car did not slow down. Soon after Elaine Hertzberg's death, Uber halted its testing program.

• 55:52Speaker 15

  Uber has temporarily suspended its driverless fleet nationwide as the NTSB, police, Uber, and the National Highway Traffic Safety Administration investigate.

• 56:01Speaker 01

  We reached out to Uber for comment. A spokesperson said that the fatal collision was indeed a tragedy, which had a significant impact on Uber and the entire industry. There would be other competitors who would shut down after similar accidents. There would also be Tesla, which by 2020 was publicly marketing a product the company called full self-driving, but which absolutely was not. Meanwhile, Waymo had slowly continued to develop its tech. Their robot taxis would be ready for riders by 2020. The team had gotten an unexpected boost from a technology that was, at the time, very little understood.

• 56:39Speaker 01

  In 2026, when most people talk about artificial intelligence, the conversation defaults to products like ChatGPT and Claude. But artificial intelligence has been a core part of driverless cars going back two decades. In the 2010s, neural net advances meant that you could now begin to feed a computer system large amounts of data and watch as its perception, prediction, and decision-making abilities improved. Here's Sebastian Thrun.

• 57:05Speaker 08

  That technology of massive data training was with us from the get-go, but has become more and more and more and more important. The surprise for all of us has been that size matters. When you put a million documents into an AI, it's fine. A hundred million is fine. But when you put a hundred billion documents into an AI, it is unbelievably smart. And that, I think, shocked everybody, myself included.

• 57:32Speaker 01

  The Google Brain team, the deep learning people, started working with the driverless car team to use training data to help the computer driver learn things like how to better predict when another car was about to suddenly switch lanes, how to more reliably spot pedestrians. Over the years, as the car drove more miles, as the team gathered more data, plugged that data into their AI systems and tweaked those systems, the engineers say the robot driver kept improving. As they tested the car in new weather conditions, they discovered problems that required hardware fixes. For instance, in Phoenix, Waymo had to design miniature wipers for their car's LiDAR sensors to deal with the dust storms and heavy rains. In 2020, Waymo finally debuts to the public in Arizona. In the years after, it'll roll out to 10 more American cities.

• 58:16Speaker 01

  A funny consequence of Waymo's long development cycle is that the public's attitude towards Silicon Valley has just really changed in that time. There's more suspicion towards Google than there was back in 2009, when the project first started. And so now, many people look at the Waymo driver with a raised eyebrow, with a question immediately on their lips.

• 58:38Speaker 01

  Chapter 5. Are you a good driver?

• 58:41Speaker 09

  All right, autonomous vehicles can now get you around Atlanta. The future of driving through Austin is here. Except, it comes without a driver.

• 58:49Speaker 16

  The fly-hailing app is now taking passengers in Miami.

• 58:52Speaker 01

  A fleet of white electric Jaguars, covered in 40 different sensors, cameras, radar, LiDAR. It's an expensive car, as much as $150,000 by some estimates. In the news stories, you see the inside, where the human driver would normally sit. There's an empty seat you're not allowed in, with a steering wheel in front of it, vestigial, that turns itself.

• 59:13Speaker 13

  Cars without drivers are here.

• 59:15Speaker 17

  Yeah, it sounds like something out of the Jetsons, but get ready, because you may look over at the car next to you and see it rolling down the street.

• 59:23Speaker 01

  The TV newscasters always use the same gee whiz tone. They can never resist the Jetsons reference. In every city, the influencers happen to record testimonials for their daily serving of clout.

• 59:34Speaker 18

  So in today's video, I'm about to take my first ever driverless car. It's with an app called Waymo. Waymo is basically driverless car Uber, where it's like ride service. You call it, go wherever you need it to go, but there's no driver.

• 59:48Speaker 19

  You guys, this is creepy. It's like I'm being driven around by a ghost person. It's a little terrifying. It is definitely.

• 59:54Speaker 01

  Rover taxis poll hilariously badly. According to J.D. Power, a data analytics firm, among people who've not ridden in one, consumer confidence is at 20%. But among people who have taken a ride, the number shoots up to 76%. It's a thing I didn't capture in this story, but when I sat in one a couple of years ago, I just found it persuasive as an experience.

• 1:00:18Speaker 20

  You know what? I'm not as nervous as I thought I was gonna be. This is actually quite relaxing.

• 1:00:23Speaker 18

  Nice gradual turn, felt very safe.

• 1:00:25Speaker 20

  You know, it was kind of freaky at first, but now it's pretty chill.

• 1:00:29Speaker 18

  It's a smooth ride, though. It wasn't driving fast, it wasn't jerking.

• 1:00:32Speaker 21

  It's driving like you always hoped your Uber driver would.

• 1:00:35Speaker 18

  So I guess that's one of the big selling.

• 1:00:36Speaker 01

  Chris Armsen, that methodical team leader, had left Google years ago, but he told me about his experience as a civilian consumer, trying a Waymo out in the world.

• 1:00:45Speaker 04

  My universal experience has been, and you can tell me if this was your experience, the first couple of minutes in the vehicle, it's, huh, that's crazy. There's nobody behind the wheel. Ooh, swimming with sharks. And then a few minutes in, it's like, okay, you know, is this just gonna drive? Is that all it does? And then, you know, 10 minutes in, people are looking at their phone.

• 1:01:10Speaker 01

  People tend to feel safe in these cars, but are they actually? So we know that the Waymo driver has now driven over 200 million real-world miles, and they've released safety data so far for the first 127 million miles. Waymo's fairly transparent. They release their crash and safety data unredacted to the public. By contrast, Tesla redacts the details of its crashes. The company says they are confidential business information. In Waymo's case, I've looked at the data. I've looked at how the company interprets it, how skeptical independent researchers interpret it. I wanted to walk through it with an autonomous vehicle reporter I trust. His name is Timothy Beeley, author of the newsletter Understanding AI. I asked him how much our picture of the Waymo

• 1:01:55Speaker 01

  safety data has been evolving.

• 1:01:58Speaker 22

  So it's been pretty consistent the last couple of years. They are scaling up and so all the numbers get bigger, like the total number of miles get bigger, the number of crashes getting bigger, but the like crashes per mile have not changed a ton. Waymo says, and I think this is correct, that it's roughly 80% safer in terms of crashes that are severe enough to trigger an airbag, crashes severe enough to cause an injury, and also crashes involving vulnerable road users like pedestrians or bicyclists.

• 1:02:28Speaker 01

  So 80% fewer airbag crashes than human drivers, and actually 90% fewer crashes that cause a serious injury. Some independent experts have small quibbles with the methodology, but broadly, they find Waymo's data credible.

• 1:02:43Speaker 01

  Timothy pointed out, there's one very important thing we don't know, the fatal crash comparison. For every 100 million miles humans drive, we cause a little over one fatal crash. The Waymo driver has driven 200 million miles without causing a fatal crash, but statistically speaking, that could still be a fluke. Some academics have suggested we need about 300 million miles to have statistical confidence. In the hundreds of millions of miles the Waymo driver has traveled, it was involved in two fatal crashes, which it did not appear to cause. Here are the details of those crashes. In one, a speeding human driver rear-ended a line of vehicles at a stoplight. There's an empty Waymo in the line of struck

• 1:03:25Speaker 01

  cars. In another crash, a Waymo was yielding for a pedestrian. It was rear-ended by a motorcycle. The motorcycle driver was then struck by a second car. That's everything.

• 1:03:38Speaker 01

  When Timothy Beeley looks at the entire safety picture, the results we have so far from this big experiment Waymo is conducting on American roads, what he sees is mainly promising.

• 1:03:49Speaker 22

  So far, it's been better than human drivers. And so far, I think the case for allowing them to continue the experiment is very strong.

• 1:03:56Speaker 01

  Which doesn't mean we shouldn't scrutinize this Waymo experiment as it continues. I find myself paying a lot of attention to Waymo crashes, which isn't hard. They make headlines. The most harrowing one recently was this January.

• 1:04:09Speaker 12

  A child near an elementary school in Santa Monica is struck by a Waymo.

• 1:04:13Speaker 23

  A child ran across the street from behind a double parked car and a Waymo hit the kid. Santa Monica police say the child, a 10 year old girl, was not hurt.

• 1:04:21Speaker 01

  The company issued a statement. Waymo said its driver had braked hard, reducing speed from 17 to under 6 miles per hour. A faster reaction, they claimed, than a human driver would have been capable of. What happened next at the accident scene actually answers a question I'd had. What does a Waymo do after a car crash, since there's no human driver to help? Waymo employs what they call human fleet response agents, human beings who can't remotely drive the cars, but who the car can ask questions to if it gets confused. In Santa Monica, the Waymo called one of those humans. The human called 911, and this is the strangest part of Waymo's statement. Apparently, the car then waited at the scene of the accident until the police dismissed it.

• 1:05:04Speaker 01

  That's what we know so far, but there's two federal agencies investigating this crash, and so we'll have a full report in the future. One problem that's not really captured in the safety data that I've seen is what I'd call troubling edge cases. You see them in videos on social media. A Waymo gets stuck at a dead stoplight, or blocks an emergency vehicle. Or, an example Timothy gave, Waymos were driving past stopped school buses in Austin.

• 1:05:28Speaker 22

  I think it's reasonable to say this is like a clear cut rule that the vehicle should follow this rule. These edge cases are still very rare, and so if it's a 1 in 10 million thing, I think it's not that big a deal, as long as they are making progress, which for most of these, I think they are.

• 1:05:41Speaker 01

  Timothy pointed to one area where Waymo's not been as transparent as he'd like. Those human response agents, some of which are based here, some in the Philippines, there's questions about what specifically they do, and about how this will all work as Waymo scales up. We asked Waymo for comment on everything you heard in this episode, especially the recent safety incidents. A spokesperson said that the data to date indicates that the Waymo driver is already making roads safer in the places where they operate, and says that Waymo can use to work with policymakers and regulators to improve its technology.

• 1:06:13Speaker 01

  That's the safety picture so far, which to me, after many months of looking at this and talking to experts, looks pretty good. As Waymo continues its rollout, other companies are quickly following behind.

• 1:06:24Speaker 24

  Amazon's new driverless taxi is launching in Las Vegas this summer, and it's expected to arrive in L.A.

• 1:06:31Speaker 01

  There's other robo-taxi companies like Amazon Zooks. Uber is back in the mix, not making technology, but partnering with these robo-taxi companies.

• 1:06:38Speaker 25

  We Ride recently struck a partnership with Uber to bring its AVs to Abu Dhabi.

• 1:06:43Speaker 24

  Another sign that

• 1:06:44Speaker 01

  And many of those early Waymo engineers are now CEOs of autonomous companies themselves. Dmitry Dolgov is actually co-CEO at Waymo, but other team members run driverless trucking companies.

• 1:06:55Speaker 25

  Got Don Burnett, founder and CEO of Kodiak AI. Don, thank you so much for joining us. It's good to see you again.

• 1:07:01Speaker 01

  Don Burnett is head of Kodiak AI, which has its technology deployed in driverless trucks in the Permian Basin.

• 1:07:06Speaker 26

  Please welcome CEO of Aurora, Chris Urmson. A big round of applause.

• 1:07:12Speaker 01

  Chris Urmson now heads Aurora, which currently has semi-trucks on Texas highways. And my personal favorite plot development, which just emerged this week.

• 1:07:20Speaker 21

  I just broke on the information that Uber founder Travis Kalanick is starting a new self-driving car company with financial backing from Uber and in partnership with Anthony Levandowski. Now, for those who've been

• 1:07:34Speaker 01

  They say there's no second acts in American lives. Somehow, both of these men seem to be on their fourth. The big picture, though, is that everywhere in America today that you see a driver, taxi, truck, food delivery, there are several companies working on the robot version, trying their best to make driver as a job start to go the way of the knocker-upper, of the lamplighter. Those knocker-uppers, by the way, they disappeared quietly. The lamplighters did not.

• 1:08:05Speaker 01

  Writer Carl Benedict Frey tells the story of the Lamplighters Union, how their strikes plunged New York City briefly into darkness to the delight of lovers and thieves. In Verveille, Belgium, the Lamplighters' strikes turned violent, ending in an attack on the local police headquarters. The army was brought in. The Lamplighters lost their fight, in part just because they were so outnumbered. But the drivers today, fighting to save their livelihoods, are a significantly bigger force.

• 1:08:32Speaker 27

  Please stand up. Everybody that's ride share, union members, or someone who drives a vehicle, stand up.

• 1:08:41Speaker 01

  4.8 million Americans strive for a living. It's one of the most common jobs we have. And these workers do not plan to surrender to the California tech companies.

• 1:08:50Speaker 28

  They're doing this because they stand to make an unfathomable amount of money if they eliminate driving jobs for working-class people.

• 1:08:58Speaker 02

  I understand it is a business. It is capitalism, but not in my city at the expense of our jobs.

• 1:09:07Speaker 01

  These drivers are represented by unions, backed by politicians. And in cities across America, blue cities, they're organizing. So far, they're winning.

• 1:09:16Speaker 27

  Humans drive this city, not machines. Labor drives this city. Keep the workers in the workforce. If it works in another city, great, have fun. Not here. Not Boston. Thank you.

• 1:09:32Speaker 01

  Next week, the fight to save a job, to save the human driver. Don't miss this one.

• 1:09:53Speaker 01

  Many thanks to PJ Vote and the entire search engine team for this story. You will hear part two right here on Freakonomics Radio very soon. Until then, take care of yourself. And if you can, someone else, too.

• 1:10:05Stephen J. Dubner?

  Freakonomics Radio is produced by Renbud Radio. You can find our entire archive on any podcast app. It's also at Freakonomics.com, where we publish transcripts and show notes. For search engine, this episode was produced by Emily Malter. The show was created by PJ Vogt and Shruti Pinnamaneni. Garrett Graham is their senior producer. Leah Reese Dennis is their executive producer. Fact checking was done by Mary Mathis, and sound design and original composition by Armin Bezarian. Their production intern is Piper Dumont. For Freakonomics Radio, this episode was produced by Dalbin Abouaji and edited by Ellen Frankman. The Freakonomics Radio Network staff also includes Augusta Chapman, Eleanor Osborne, Elsa Hernandez, Gabriel Roth, Elaria Montenacort, Jasmine Klinger,

• 1:10:48Stephen J. Dubner?

  Jeremy Johnston, Teo Jacobs, and Zach Lipinski. Our theme song is "Mr. Fortune" by The Hitchhikers, and our composer is Luis Guerra. As always, thanks for listening.

• 1:11:03Stephen J. Dubner?

  Is it possible that you were really stoned on painkillers in that first Waymo ride?

• 1:11:08Speaker 01

  I mean, I wasn't stoned on painkillers, and I don't think I was stoned at all. I think I really had a sense of normal technological awe.

• 1:11:21Speaker 10

  The Freakonomics Radio Network, the hidden side of everything.