Stürm und drang —

Quest to dominate self-driving cars is at the heart of Waymo v. Uber trial

Judge: case is about trade secrets, “not whether or not Uber is an evil corporation.”

A customized Waymo minivan being shown off at the 2017 North American International Auto Show in Detroit.
Enlarge / A customized Waymo minivan being shown off at the 2017 North American International Auto Show in Detroit.

This story has everything: a cutthroat startup using cloak-and-dagger tactics going up against a corporate giant, a now-dismissed star engineer accused of stealing thousands of crucial files, cutting-edge Silicon Valley technology, and a whipsmart, no-nonsense, tech-savvy judge.

This year's hottest tech-related trial is Waymo v. Uber. At 7:30am PT Monday, what began as a trade secrets lawsuit back in February 2017 is set to open in federal court in San Francisco, just blocks from Uber headquarters.

On one side is Waymo, the self-driving car division of Alphabet, Google's parent company. Its lawyers accused Uber of effectively stealing its secret sauce when the ridesharing company hired away a top engineer who had taken 14,000 internal files without authorization before leaving Google.

On the other side is Uber, which maintains that it did nothing wrong when it hired Anthony Levandowski in August 2016 and paid an eye-popping $680 million for his startup, Otto, which had only existed for a few months. Uber maintains that it hasn't improperly benefited from what Levandowski took. In fact, since this case began, Uber even fired Levandowski because he would not comply with subpoenas.

What precise trade secrets Uber is accused of taking remains publicly unknown. For Waymo, showing that what Uber acquired was indeed a trade secret and that it was misappropriated isn't a slam dunk. After all, while trade secrets are, as the moniker suggests, secret, they aren't as legally protected as patents.

"Waymo will need to prove that Levandowski stole its information, that Uber knew it, and that Uber incorporated [the information] into its self-driving car technology," Russell Beck, a veteran trade secrets lawyer, told Ars. He also noted that usually such lawsuits settle rather than go to trial.

"Uber will need to show that, even if there was some wrongful conduct, it is not using Waymo's technology and, therefore, no harm, no foul."

Still, if Waymo can prove that Uber is liable and improperly used its trade secrets, Uber could be on the hook for billions of dollars—potentially a sizeable blow to a company that has already lost billions as it aggressively expands globally. If Uber loses at trial and also loses on appeal, it may pose existential problems for a company that is already being heavily scrutinized.

High-stakes gamble

Separately, Uber continues to be under the microscope for its questionable internal practices. So far, the company has been criticized for tolerating sexual harassment and thwarting regulators, and inquiries have been made into Uber about potential criminal behavior. Its previous CEO Travis Kalanick also had a penchant for secrecy and power during his tenure.

In 2016, Kalanick characterized Uber's attempt to at least stay atop the autonomous vehicle space as critical to the company's future.

"If we are not tied for first, then the person who is in first, or the entity that's in first, then rolls out a ridesharing network that is far cheaper or far higher-quality than Uber's, then Uber is no longer a thing," he told Business Insider.

In a July 2017 deposition, Kalanick also said: "in order for Uber to exist in the future, we will likely need to be a leader in the autonomous vehicle space."

In short, Uber has more to lose.

Then-CEO Travis Kalanick gestures as he speaks at an event in New Delhi on December 16, 2016.
Enlarge / Then-CEO Travis Kalanick gestures as he speaks at an event in New Delhi on December 16, 2016.
MONEY SHARMA/AFP/Getty Images

However, if Uber can stave off the lawsuit and win on appeal, then its billion-dollar bet last year on self-driving cars will likely let it stay in the game. After all, Uber currently has what Waymo doesn't have: a vast ridesharing network and the software and engineers to go with it. Waymo may be ahead in AV hardware and software, but Uber is dominant in rideshare software.

"Uber will need to attack the validity of the trade secret status of the information but, most importantly, show that [Uber] had no knowledge of what Levandowski was doing and that [Uber] did not acquire or use the information," Elizabeth Rowe emailed Ars. She is a professor at the University of Florida and expert in trade secrets law.

"As a matter of strategy," she added, Uber "may try to distance [itself] from him as much as possible."

Whoever wins this case, it's clear that a dominant autonomous rideshare company will emerge in the coming years: it may be Waymo, Uber, or yet another company, like GM.

"While Waymo and Uber duke it out in court, 'traditional' players like GM, with their manufacturing and marketing muscle, could end up as the winners," Raj Kumar, the co-director of the General Motors-Carnegie Mellon Vehicular Information Technology Collaborative Research Lab, emailed Ars.

“Waymo and Uber, for all their technological prowess, lack the underlying platform, and all their alliances with carmakers seem to have been pretty weak and toothless. They could have been just window dressing for carmakers to know what Waymo and Uber are capable of in the AV space."

Origin story

The modern era of autonomous driving really began in 2004, during the DARPA Grand Challenge. This was a Department of Defense-funded competition with a cash prize of $1 million for the first team to build a vehicle that could travel 150 miles autonomously in the desert of southeastern California.

None of the teams finished the challenge that year. The team that finished the farthest, "Red Team" from Carnegie Mellon University, barely exceeded seven miles of the course. One top member of the squad was Sebastian Thrun, then a CMU professor. One of Thrun's competitors was Levandowski, then 24-years-old, who had recently gotten his master's degree in industrial engineering from UC Berkeley.

The young engineer had an entirely different strategy at the Grand Challenge: rather than try to outfit a traditional four-wheeled car, why not do it on a motorcycle, which would be nimbler and cheaper? His team's creation, Ghostrider, was born.

(The motorcycle's name may have come from a a phenomenon born in nearby Oakland called "ghostriding," in which people would dance on or near a car rolling slowly with no one behind the wheel. The term was immortalized in the 2006 Mistah F.A.B. song "Ghost Ride It.")

Anthony Levandowski (far left) seen here preparing for the DARPA Grand Challenge in 2004.
Enlarge / Anthony Levandowski (far left) seen here preparing for the DARPA Grand Challenge in 2004.
Kim Kulish/Corbis via Getty Images

The following year, 2005, Thrun was at Stanford University, where he and his team built up "Stanley," a Volkswagen SUV. Both Stanley and the Ghostrider competed against one another, and Stanley ended up winning. It completed the 132 mile-long desert course in just under seven hours. Thrun and his team collected the prize, which had been doubled to $2 million.

Among the attendees of the 2005 DARPA Grand Challenge were Google co-founders Sergey Brin and Larry Page, who turned up to the event in hats and sunglasses, lest they be recognized.

"Larry has been a believer in this technology for much longer than I even knew," Thrun told CNBC in 2017. "And so was Sergey. And they really want to understand what's going on."

In the wake of the 2005 event, while still at Stanford, Thrun began pursuing research into autonomous vehicles with even more vigor. Likely in 2006, he began work on VuTool, an imaging startup that used technology from the research on Stanley. One of Thrun's colleagues at VuTool was Levandowski—the entire startup was reportedly acquired by Google around March 2007.

VuTool became critical in the early efforts of Google Street View. According to the Wall Street Journal, VuTool's tech was not only cheaper, but superior to what Google had made in-house.

However, even after VuTool was acquired and Levandowski became a Google employee, he went ahead and founded a new company, 510 Systems, on May 11, 2007. (510 refers to the area code for Berkeley and the surrounding region.)

DARPA held no event in 2006, but in November 2007, it decided to up the ante once more. This time, rather than having the teams compete in a rural area, it would invite teams for an "Urban Challenge" held on an old Air Force base in Victorville, California—100 miles east of Los Angeles. That base, with its streets and buildings, would simulate a small town. Teams had to complete a 60-mile course, navigate intersections, and obey all California traffic laws.

There, Thrun and his Stanford colleagues entered a converted Volkswagen Passat named "Junior." In the end, they took second place and a $1 million prize.

Channel Ars Technica