A LITTLE AFTER NINE IN the morning of a cool Friday in April 2017, Anthony Levandowski sat down where so many of his colleagues and friends had predicted he would land himself: in a conference room surrounded by lawyers, being grilled about his starring role in the first great battle of a world he had helped create.
If the blinding morning sun hadn’t been coming through the window of the twenty-second-floor office in downtown San Francisco, Levandowski would have been able to see the Bay Bridge. Every day, 260,000 vehicles used the 8.4-mile span to cross the bay that divided the city from Oakland, Berkeley, and the rest of its East Bay neighbors. By six in the morning, the mass of cars, trucks, vans, and motorcycles waiting to pay the ever increasing toll and funnel onto the crossing created a mile-long parking lot. On days when someone crashed on the bridge, the resulting extra congestion could cripple the region’s road network. Like eighteenth-century urbanites emptying chamber pots from upper story windows, it was a quotidian sort of insanity, excused by entrenchment and a lack of better options.
Attorney David Perlson, of the white shoe law firm Quinn Emanuel Urquhart & Sullivan, began the deposition. “Where do you work currently?”
“I work at Uber,” Levandowski said.
Six feet six inches tall and slim, with a head of dark hair that was starting to recede, Levandowski wore a blue suit for the occasion, no tie. Apart from the black sneakers, it was a rare change from the standard Silicon Valley engineer look he embraced: jeans and whatever T-shirt was on top of the dresser drawer that morning.
“Okay,” Perlson said. “And what’s your position there?”
“I’m vice president of engineering.”
“What are your responsibilities as vice president of engineering?”
Here, at the direction of his lawyer, Levandowski read from a piece of paper on the table in front of him.
“On the advice and direction of my counsel, I respectfully decline to answer,” Levandowski said. “And I assert the rights guaranteed to me under the Fifth Amendment of the Constitution of the United States.”
“How long have you worked at Uber?”
“On the advice and direction of my counsel, I respectfully decline to answer. And I assert the rights guaranteed to me under the Fifth Amendment of the Constitution of the United States.”
Over the following six hours, Levandowski declined to answer one question after another, questions that in their one-sidedness built a damning narrative.
“When you worked at Google, you received tens of millions of dollars in compensation from Google, is that accurate?”
“You and Uber discussed how you would form a new company while you were employed by Google?”
“You and Uber discussed that your new company would eventually be acquired by Uber while you were still employed at Google?”
“That new company eventually became Otto, correct?”
“While you were still employed by Google, you recruited engineers to join your new company so that your new company could replicate Google’s Lidar technology, correct?”
“You took over fourteen thousand confidential files from Google prior to your departure from Google, correct?”
“You took the fourteen thousand documents from Google so that you could get—so that you could more quickly replicate Google’s technology at Otto, correct?”
“Mr. Levandowski, your use of the fourteen thousand confidential documents you took from Google allowed you to sell Otto to Uber for over $680 million in just a few months?”
Again and again and again, Levandowski gave his carefully scripted nonanswer, citing his Fifth Amendment rights.
Officially speaking, Levandowski was just one of many witnesses being deposed in the run-up to Waymo v. Uber, a legal brawl between two corporate giants. Waymo had started life as a Google project called Chauffeur, and was now its own company under the umbrella of Google’s parent company, Alphabet. Uber was the enormously valuable ridehailing company that had thrown the world of urban transportation into chaos since its founding in 2009. Both were racing to create and deploy cars that could drive themselves.
Their fight centered on the thirty-seven-year-old Levandowski, who had spent nine years at Google before moving to Uber. In Waymo’s telling, on December 14, 2015, Levandowski downloaded more than fourteen thousand technical files from its servers onto his laptop, many of them describing the inner workings of its all-important Lidar laser vision system. He connected an external hard drive into the computer for eight hours, then installed a new operating system to wipe away evidence of the downloads. He quit six weeks later and founded Otto, a company dedicated to developing self-driving trucks. After a few months, Uber acquired Otto for a reported $680 million—an astounding figure for such a young company—and put Levandowski in charge of its own autonomous driving project.
Under Levandowski’s direction, Waymo alleged, Uber’s engineers used those files to accelerate their technical progress and play catchup, having started their research only in 2015, six years after Google. That, Waymo insinuated, was why Uber had been able to send robotic trucks along the highways of Colorado and Nevada, how it was using robotic cars to move people around Pittsburgh. Those vehicles still had people behind the wheel, but it was only a matter of time—time better counted in months than years—before the flesh-and-blood backups were no longer necessary.
Uber said that nothing Levandowski may have taken made its way into its work.
If Waymo’s phalanx of lawyers convinced the jury that Uber had cheated to get ahead, Uber could be forced to put its autonomous driving efforts on ice, or maybe the scrap heap. And that wouldn’t be just a hit to the balance sheet. It would be an existential crisis. Driverless cars would be safer and cheaper than human-driven ones, and any service that provided them would dominate the market, said Uber CEO Travis Kalanick. “In order for Uber to exist in the future, we will likely need to be a leader in the AV, autonomous vehicle, space.”
Kalanick was right. Robots will drive the future. By the start of the Waymo v. Uber trial in February 2018, fleets of autonomous vehicles were roaming the streets of Silicon Valley, San Francisco, Pittsburgh, Phoenix, Detroit, Boston, Munich, and Singapore—to name a few. Tesla, Cadillac, BMW, Audi, Mercedes-Benz, Nissan, and other automakers were selling cars that could pilot themselves on the highway. Along with Google and Uber, Ford, General Motors, and others were working on fully driverless cars that wouldn’t need steering wheels or pedals. Dozens of companies, from the world’s largest corporations to the smallest startups, were crowding into a technology whose upside flirted with utopianism. The average American worker spent nearly an hour driving to and from work every day; driverless technology would turn that chore into free time. Robots that never get drunk, tired, angry, or distracted promised to drastically reduce crashes, more than 90 percent of which result from human error. Those crashes kill about forty thousand Americans every year. Globally, the annual death toll is well over a million.
Uber and Waymo executives sang sweet songs about ending road deaths, but they weren’t in court fighting over who got to save more lives. They went to war because each wanted to claim a dominant share of a market predicted to be worth $42 billion in 2025 and $77 billion in 2035, when 12 million new robo-cars would hit the road annually. By 2050, autonomous driving tech could add $7 trillion to the world’s economy, all of it for the taking by anyone who could make it safer to get around, cheaper to move goods, and way more relaxing to be stuck in traffic.
That was the near term. The advent of the personal car shaped the world’s cities, suburbs, and rural areas over the past century. It created cultures. It inspired art; it was art. It helped create and define the middle class. Questions remained about how autonomous cars would be tested, certified, insured, and operated. But these were details. The shift away from human driving promised to be as influential as the car itself, if not more so. It offered the opportunity to remake cities, to correct the mistakes of the past.
Driverless cars would be shared, and they’d be cheaper than today’s taxis or Ubers. They wouldn’t need to take up precious urban space for parking, instead driving themselves to lots in less dense areas. They’d run on electricity instead of gasoline, reducing pollution and helping balance the power grid. They’d boost productivity. Many more effects were hard to anticipate. Just as the smartphone begat an app ecosystem, including a ridehailing market dominated by Uber, robotic driving could create entirely new industries.
Critics and skeptics feared the tech would encourage suburban sprawl, since people wouldn’t mind long commutes if they could work, sleep, or relax on the road. The promise of smarter cars could sap officials’ interest in funding reliable, equitable public transit. Self-driving cars could ruin the fun for people who like being behind the wheel, and multiply cybersecurity risks for everyone, giving malevolent hackers a juicy new target. And they were poised to eliminate, over the coming decades, the jobs of the 4 million Americans who made a living by driving.
The reality, though was inescapable. The age of autonomous vehicles was coming, and—like sails, steam, combustion engines, and the physics of flight—the technologies propelling it along would turn the world on its head.
By the time Anthony Levandowski stepped into the conference room for his deposition, he had seen his reputation, and possibly his career, transformed into a smoking ruin. The judge called Waymo’s account “one of the strongest records I’ve seen for a long time of anybody doing something that bad.” In May, he took the unusual step of recommending that the Department of Justice investigate criminal charges. Uber fired Levandowski a week later.
Pleading the Fifth—Levandowski would repeat those two sentences 387 times in that day’s deposition—may have protected him legally, but it also meant that at the moment his critics were loudest, no one spoke in his defense. No one would say that he had been there at the beginning of Google’s self-driving research, that he had done more than perhaps anyone else to bring this technological revolution to the brink of reality. But at the end of that Friday in April, with twenty minutes left in a six-hour deposition, Uber’s lawyer asked him how he had heard about something called the DARPA Grand Challenge, while Levandowski was a graduate student at the University of California, Berkeley. Since the topic had little to do with the facts of the case, his lawyer let him respond. The answer took him back to a 2003 conversation with his mother.
“My mom knew how much I loved robots and that I loved making things. And she gave me a call when she found out about this competition sponsored by the Defense Department,” Levandowski said. “When I saw it, I couldn’t resist.
“It was a race from LA to Vegas, across the desert,” he went on. “The goal was to release a vehicle into the world on its own without any remote control or assistance” and have it go from start to finish, all on its own. His entry, he said, was a motorcycle called Ghostrider. “I built a substantial portion of it myself, but I also created a team to help me,” Levandowski said. “It was, frankly, a pretty crazy idea.” The self-riding motorcycle didn’t reach the other side of the desert, but it did make its way across the country when America’s great museum came calling. “I donated it to the Smithsonian, where it is today.”
A few questions later, the lawyers were back to the more recent past and Levandowski was back to the Fifth Amendment, back to a defensive silence. But his contribution to the annals of technological history didn’t end with landing a robotic motorcycle in the Smithsonian. That was simply where it began.