It’s hard to know where this all ends. Will the winner be the company that develops a vehicle capable of driving along a preprogrammed route, in ideal weather conditions, in certain cities? Or will it be the vehicle that can drive itself anywhere, rain or snow? Is it enough to build a vehicle that drives itself most of the time, or is a truly driverless car one in which the passenger can safely take a nap? Does it count if a company produces just a handful of expensive prototype vehicles, rather than something that can be mass-produced and sold by the tens of thousands?
Waymo has set a narrow target. It is aiming for Level 4 autonomy, an official classification for a vehicle that is capable of driving itself, with no human behind the wheel, in most environments and road conditions. It believes that nothing short of Level 4 counts as autonomous, and that bypassing Level 3 (a lower classification, in which some human attention is still required) is necessary to keep people safe on the roads. Humans, the company says, can easily lose concentration while driving with Level 3 autonomy, even though they often need to take over at a moment’s notice.
That is also a self-serving goal. By most estimations, Waymo is closer than any other company to Level 4 autonomy. Its test vehicles, which have been on the road for eight years, have completed more than 3.5 million autonomous miles, far more than any of its competitors. It has driven another 2.5 billion miles in “Carcraft,” a simulated, virtual-reality environment that allows the company to run and rerun various situations millions of times a day, and incorporate the results into its real-world vehicles. And its access to Google’s top-tier engineering talent and cutting-edge technical infrastructure has made it a formidable competitor.
Waymo’s advantage was apparent on Monday, during my test ride in the company’s self-driving minivan. Unlike other driverless demos I’ve encountered, there was no human keeping watch at the wheel. To start the trip, a colleague in the back seat pressed a large blue button marked “start ride,” and the van took off on its own. In a 10-minute drive around Castle, it easily handled a series of different on-road situations and obstacles. It encircled a roundabout, waited patiently for pedestrians to cross the street, and dodged traffic cones and bicyclists with ease. It felt slick and polished, and gave the sense that it might be ready for public consumption very soon.
Granted, this was a carefully planned test loop in a controlled environment. Autonomous driving on public roads is a much harder challenge, and Waymo is still figuring out how best to deploy its technology to a wider market. The company declined to put a date on when it might release self-driving cars to the general public, and Mr. Krafcik spoke only in generalities about its plans, saying that it would focus on ride-hailing and autonomous trucking as possible early business models.
“We’re really close,” he said.
But Waymo’s road to autonomy has not been entirely smooth. It has lost some engineering talent to well-funded competitors. Its executives have been distracted by a prominent lawsuit against Uber, in which Waymo accused Uber of stealing trade secrets worth more than $2 billion. And in an early-access test program in Phoenix, its vehicles reportedly struggled to make left turns, one of many real-world obstacles that have to be overcome.
And the competition is heating up. Lyft and Drive.ai, a self-driving car start-up, are teaming up to offer a pilot program for driverless ride-hailing in San Francisco. Uber has been developing its own driverless fleet, and it is testing vehicles on public roads in Arizona, Pittsburgh and Toronto. Detroit is getting up to speed, too — General Motors, which spent $1 billion to acquire Cruise Automation last year, recently announced that it has a production-ready autonomous vehicle, once the software and regulatory kinks are worked out, and Ford, which has partnered with Argo AI, an artificial intelligence company based in Pittsburgh, plans to put Level 4 autonomous vehicles on the road by 2021.
Waymo’s victory against this army of well-funded competitors is not guaranteed. But it won’t be easy to beat, especially if it gets to define what winning means.
Continue reading the main story