America Is About to See Way More Driverless Cars
Robotaxis are now picking up riders in L.A. and Houston—and facing a whole new set of challenges.
The future of driverless cars in America is a promotional booth with a surfboard and a movie director’s clapboard. Robotaxis have officially arrived in Los Angeles, and last week, residents lined up in Santa Monica’s main promenade to get a smartphone code needed to ride them. For now, the cars, from the Alphabet-owned start-up Waymo, won’t leave the tame streets of Santa Monica. But in the coming months, they’ll embark on a multi-month “tour” of the city, heading to West Hollywood, downtown L.A., and several other neighborhoods.
For the past decade, the two leading robotaxi companies, Waymo and Cruise, have been focused primarily on San Francisco and Phoenix, where they both already take paid passengers. But now they are expanding into new cities, adding millions more potential riders (and bystanders) into the mix. Last week, Cruise, a subsidiary of General Motors, launched its robotaxi service in Houston, and will also soon do the same in Dallas. In addition to L.A., where Cruise is testing and reportedly will soon expand, Waymo is kicking off in Austin, where Cruise already takes passengers.
In San Francisco in particular, the cars have jammed up in ways both silly and serious. Local protesters, fed up with the technology, have put cones on top of the cars to confuse their navigation system. But now the vehicles will face new challenges: As they move beyond their hometowns, their systems will be tested on new kinds of streets, with different driving cultures and different rules. They’ll have to drive in notorious L.A. traffic and notorious Houston traffic. Robotaxis haven’t had it easy in San Francisco, but the race to go national might still be bumpy.
If self-driving cars can handle San Francisco, they should ideally be able to handle anywhere. The city is “by far the toughest environment anywhere in the world to test such technology due to topography and complex street geometries,” Rahul Jain, a professor at the University of Southern California and the director of its Center for Autonomy and Artificial Intelligence, told me over email. Cruise says it chose to start there in part because of these challenges, which the company argues makes it a good place to build this technology. In San Francisco, I’ve ridden in cars made by each company, and in both instances, the cars navigated the streets with impressive ease. L.A. should be easier, but it poses new challenges in terms of traffic and driver aggression. Having lived and driven in both cities, I can confirm that L.A. drivers have more of a knack for speeding and changing lanes in tight highway traffic. “You have to be a little bit wild to be able to drive here,” Jain joked.
In L.A., Waymo cars won’t go on the highway, but the rollout still will not be easy. Both companies test their cars for many hours before passengers can start riding, but that doesn’t mean it’ll be perfect when the vehicles go fully operational; after all, they were likewise tested in San Francisco for many years. Both Waymo and Cruise may soon learn the ways that their computer-vision systems are and aren’t generalizable between cities, says Missy Cummings, the director of George Mason University’s Autonomy and Robotics Center: “They may have to retrain a bunch of their neural nets.”
As each other’s biggest competitor, both Cruise and Waymo appear eager to expand and signal that they are one step closer to robotaxi supremacy. I asked Cummings whether she thought these two companies should be taking passengers in Los Angeles. For Cruise, she offered a blunt no. She was softer when it came to Waymo: “I think it makes sense for them.” Cruise in particular has had a rough past few months in San Francisco—a Cruise car drove into wet cement, another collided with a fire truck, and a pedestrian was pinned under a vehicle earlier this month after it was initially hit by a human-driven car. (Cruise tweeted its account of the crash: Their vehicle “braked aggressively to minimize the impact,” and “at the request of the police the AV was kept in place.”) In August, Cruise was asked by California’s DMV to halve its operations in the city.
“Everything I see indicates Cruise has more issues with road safety, but it’s difficult to be sure, because the companies are so opaque with their data,” Phil Koopman, a professor at Carnegie Mellon University who specializes in autonomous-vehicle safety, told me. Cruise and Waymo are required to make some numbers public to regulators, reporting any accidents and, in California, disclosing the number of miles driven and any incidents in which a human had to intervene. The companies also release various safety research on their respective websites. But gaps remain: Researchers don’t have precise information about how many robotaxis are operating and where, for example. Nor do they have video footage of every crash.
A spokesperson for Cruise defended its safety record via email, pointing to the company’s own research indicating that its driverless cars are involved in 65 percent fewer collisions than human drivers are in a comparable driving environment. “Cruise drove 10x more miles than any other autonomous vehicle company in San Francisco last year,” the spokesperson added.
What that means is that we’ll likely see more of these kinds of incidents as the robotaxis, and Cruise in particular, spread beyond San Francisco. The most troubling prospect is that a new era in which self-driving cars are operating in even more cities may bring issues experts can’t predict. Noah Goodall, a senior research scientist at the Virginia Transportation Research Council, told me he was surprised that the vehicles have issues navigating emergency services. But that’s par for the course when you’re building something new. “When you’re creating a technology that’s safer, you’ll create other risks that are new that you didn’t have before,” he explained.
None of this is stopping Cruise or Waymo from slowing down. The two companies are in an arms race to advance their self-driving cars, a competition that is also between two industries. Waymo, which began as Google’s self-driving car project and is owned by Alphabet, is as Silicon Valley as start-ups come. Cruise, meanwhile, is owned by General Motors, a legacy automaker that hasn’t historically been anything like a tech company and is far less cash-flush.
Now driverless cars and their promise to reduce America’s road deaths is collapsing some of the differences between car companies and tech companies. “What Silicon Valley learned is the car business is a very, very sophisticated business, and it’s not really easy to design, engineer, manufacture, distribute a vehicle with critical safety systems and so many parts in it,” Lawrence Burns, a former GM executive and former adviser to Waymo, told me over email. “What the auto industry learned from this is that there’s extraordinary talent and capability in Silicon Valley for digital technology, software and experience design applied to the future of transportation.”
The robotaxi race is just one front on which Americans are seeing these changes up close. Electric cars are now so software-enabled that they are often described as “smartphones on wheels.” And many new cars for sale, including those from GM, are stocked with self-driving and autopilot features that far exceed what was available even a few ago. Los Angeles and Houston will bear witness to the next era of robotaxis, to whatever accidents and missteps they will surely make. They already are. Before Cruise even officially launched in Houston, three of its vehicles reportedly stalled at the same intersection, locking up traffic.