Inside the battle for the future of a technology that could really, truly change the world
The stakes are impossibly high. Self-driving cars are arguably the great technological promise of the 21st century.
They are in that rare class of technology that might actually change the world. And not just in the way everyone in Silicon Valley talks about changing the world—but really, fundamentally change it. So much so that their dramatic life-saving potential is almost secondary to the other economic, cultural and aesthetic transformations they would cause.
Those who aren’t able to drive themselves today—people who are blind, for example—would be granted a new level of transportation freedom. Mass adoption of self-driving cars would create and destroy entire industries, alter the way people work and move through cities, and change the way those cities are designed and connected.
To build the technology that prompts all this change is to be in an enormous position of power.
That’s why the race to bring self-driving cars to the masses is so intense. It’s also what makes this particular competition echo other transformative moments in technological history—going all the way back to the Railroad Wars, at least. (Incidentally, there was a different kind of driverless car back then.)
“The Wright brothers jump into my brain immediately,” John Leonard, an engineering professor at M.I.T., told me in 2015. “But maybe it’s kind of like a decentralized space race. Like Sputnik, but between the traditional car companies and their suppliers versus tech companies and their startups.”
There’s a lot of money at stake. A lot a lot. We’re talking billions of dollars per year in potential profits, maybe more. All of the major players know this. For some companies, it is a fight to the death. Each one intends to come out on top.
* * *
Waymo (formerly Google)
When Google (now Alphabet) launched its self-driving car program in 2009, it had no competition to speak of. Culturally, the idea of a self-driving car was novel. Even the flying cars in 20th-century science fiction tended to have human drivers. So when Google began to go public with information about the project, in 2010, its level of seriousness about the effort wasn’t yet clear.
“Some of these things will turn out to be wildly successful, and others will just fade away,” one investor told The Los Angeles Times at the time, referring to Google’s suite of unusual projects.
Wild success still isn’t a guarantee, but it’s now obvious that Google—which has since spun off its self-driving-car unit into a company called Waymo—is deeply invested in the work it’s doing. Its test fleet is now on public roads in four states: California (since 2009), Texas (2015), Arizona (2016) and Washington state (2016).
“We’ve self-driven more than 2 million miles mostly on city streets,” Waymo says on its website. “That’s the equivalent of over 300 years of human driving experience, considering the hours we’ve spent on the road. This builds on the 1 billion miles we’ve driven in simulation in just 2016.”
All that driving and a near-perfect safety record—a reputation that has undoubtedly helped buoy the public’s perception of self-driving vehicles.
Uber catapulted itself into the self-driving car space in truly Uberesque fashion: With a scandal. In 2015, the ride-sharing giant hired an entire department away from Carnegie Mellon—some 40 robotics experts and engineers, including several top experts in autonomous-driving systems.
Since then, Uber’s commitment to the future of self-driving cars has only intensified. (Consider the business incentive of eliminating human drivers, who get a cut of every ride they give.) In 2016, Uber began testing its self-driving vehicles on public roads in Pittsburgh, it doubled down its own proprietary street-mapping system—ostensibly to reduce reliance on competitors like Google and Apple—and poached Google’s top mapping expert to do so. Uber also acquired a fledgling self-driving truck company, Otto, for $680 million in 2016—but more on that in a minute.
Given the talents of its employees and how much venture capital the company has on hand, Uber has emerged as a formidable player in the emerging self-driving car industry. Yet, Uber continues to be plagued by controversies.
After a dustup over Uber’s refusal to seek permits for its self-driving cars in California in late 2016, the company changed course and applied for a state testing permit. In February 2017, Waymo filed a federal lawsuit claiming a former Google engineer had stolen self-driving car secrets before leaving the company to found Otto. Waymo says when Uber acquired Otto, the former Google engineer used the information he allegedly stole to help build a circuit board for Uber’s self-driving car systems. The legal battle is poised to be the first major intellectual-property fight of the driverless car era.
Apple remains one of the more mysterious and intriguing players in the self-driving car game. On one hand, Apple can’t afford not to pursue this emerging technology if many of its major competitors are. On the other, Apple? A car company? To be fair, though, that’s what people said of Google in 2010. And not all self-driving car companies will manufacturer vehicles themselves; some will just lease out the self-driving software for auto manufacturers.
For years, it was rumored that Apple had a secret self-driving car project in the works. But there were also reports the project— which according to The Wall Street Journal had hundreds of dedicated employees as of 2016—was plagued by organizational and managerial problems. It wasn’t until December of 2016 Apple officially made it known it is working in some capacity on self-driving cars, via a letter to the National Highway Traffic Safety Administration.
“The company is investing heavily in the study of machine learning and automation,” wrote Steve Kenner, Apple’s director of product integrity, “and is excited about the potential of automated systems in many areas, including transportation.”
Other than that, however, Apple has remained characteristically secretive about its work.
Tesla wants to bring driverless cars to the market, but it has a markedly different approach than Waymo, which may be its biggest competition. While Google wants to build fully self-driving systems from the ground up, its critics say this will take too long. In the interest of making everybody safer sooner, Tesla is adding increasingly autonomous systems, bit by bit, to its existing high-end vehicles.
But there’s a big debate over which method—fully autonomous versus incrementally autonomous—is actually better for public safety.
Tesla’s CEO, Elon Musk, has said it’s “morally reprehensible” to wait until the technology is advanced enough for complete autonomy. Yet, critics of the Tesla approach say here-and-there semi-autonomous features may present too much of a gray area for today’s drivers to safely navigate. The marketing around Autopilot, the current Tesla system, has arguably left people with the impression that Tesla’s cars are more autonomous than they really are. The very name, Autopilot, certainly suggests it might be okay for human drivers to stop paying attention.
This concern came up again in the spring of 2016, when a Tesla driver who was using the Autopilot feature died in a car accident. At the time, Tesla’s Autopilot feature was in beta mode, meaning the drivers who tested it on public roads were required to acknowledge any risks involved. Federal investigators eventually concluded Autopilot was not to blame in the fatal crash.
Tesla already claims on its website all of its vehicles “have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver.” This is, at best, slightly misleading. Tesla’s hardware may eventually allow for a “full self-driving” system, but it’s definitely not there yet.
Whichever approach to building a truly autonomous car is the right one, Tesla’s sense of urgency is helping to quicken the pace of competition in the driverless-car space.
Like Tesla, several legacy automakers are announcing their entry into the driverless-car space with incremental assisted-driving systems. This approach makes sense for them: After all, they already manufacture cars that people can go and buy—something that isn’t true of Apple, Google, or Uber—which means one of the best hopes for legacy carmakers to stay in business is to evolve now rather than attempting to play catchup later (which they may still have to do).
But some legacy companies have gone farther than others. While nearly every major automaker pays lip service to the importance of developing autonomous vehicles, only some have backed up their talk with action.
Volvo stands out among the more committed, for instance. In a project Volvo is calling Drive Me, the automaker will put a fleet of 100 driverless cars on the highways in Sweden. (As with tests by Google and Uber on public roads in the United States, humans will sit behind the wheel, ready to take control of the vehicles if needed.) In March 2017, Toyota unveiled its first self-driving car prototype. The car came out of Toyota’s artificial intelligence research institute, which it launched with a $1 billion investment in 2015.
There are also some partnerships between tech firms and automakers. Chrysler and Google announced in May 2016 they would team up to make a driverless minivan, while Volvo and Uber announced a partnership in August 2016.
We should expect to see more startups in the self-driving car space in the years to come. One example is Drive.ai, which launched in August 2016 and is creating deep-learning software for driverless cars.
There will be others. Chris Urmson, the longtime head of Google’s driverless car initiative, left the company in August 2016, at a time when the project seemed to be shedding several key players. In December 2016, the technology-focused news website Recode reported Urmson is starting his own self-driving-cars venture.
Technology history tells us the first company to build a technology is not always the company that ends up making a windfall. That may well be the case in the realm of autonomous vehicles.
There are many uncertainties in all this, but one thing is clear: The cultural space occupied by the automobile is undergoing rapid, radical transformation. There are sure to be big winners and losers along the way.