San Francisco Takes First Step Towards The Grim, Autonomous Future Of Transportation

For the past decade you could see and hear them crawling through San Francisco, emitting a faintly cetacean purr: hundreds of vehicles owned by the self-driving car companies Cruise (a subsidiary of GM) and Waymo (a subsidiary of Google), putt-putting around the city. The fleet's cars were often staffed by a human driver to keep them on course, and always spangled with an array of cameras pointing in every direction. These cars of course got into all sorts of whoopsies and goofs—self-driving cars are not yet much good at driving themselves—but the mundanity of seeing them all around the city every day was the most ominous part. Both the unseen San Francisco officials letting a pair of private companies befoul public space and the droopy contract drivers helping those companies obviate human labor, like a prisoner helping the hangman with the noose, symbolized a deep resignation. If the machines are going to win, why resist ceding control of the street?

San Francisco is the ideal testing ground for this sort of technology, as it's very small, close to Silicon Valley, ruled by a small group of people who have a long track record of letting big tech do whatever; the former San Francisco mayor and current ghost haunting city hall Willie Brown is being paid by Cruise, for instance. The city is also hilly and dense enough to serve as a strong proof of concept for the technology. San Francisco has allowed the two companies to use its streets as testing labs since 2013, gathering geospatial data, figuring out how to stop repeatedly hitting the dead-end on 15th Avenue in the Presidio, and learning how not to hit people. (They didn't train any robots to deploy the proper, exculpatory legalese to blame a dog for getting killed by a Waymo, and still have to use people for stuff like that.) The companies eventually expanded services to offer a small number of rides under strict conditions. The goal was always to continue improving until such time that the technology became sufficiently advanced, and state officials sufficiently pliable, to allow these cars to rove the streets and serve as an unsleeping 24-hour taxi fleet.

That time has now come. On Thursday night, the California Public Utility Commission voted 3-1 in favor of allowing the two companies to operate robotaxis in San Francisco.

The vote came on the heels of several days of contentious public debate between a local activist faction and a group of mostly astroturfed safety advocates. The case for robotaxis—or anyway the one being made to the public, which should not be confused for the real one, which is more money for a tiny investor class—is that they will make the streets of San Francisco, and soon ideally The World, safer by taking fallible human drivers off the road in favor of robots that don't fall asleep or eat drugs or compose randy DMs while driving down Van Ness at 1:47 a.m. If you think about it for like 30 seconds, this case makes a good deal of sense. Thirty-nine people were killed by cars last year in San Francisco, and traffic fatalities are rising across the country. The companies are also pushing the notion that robotaxis will be a transit panacea for people with disabilities, a claim that many advocates have greeted with serious skepticism.

Automated vehicles are safer, though it's not clear to what degree because both companies have refused to make their data public, much to the frustration of city officials. What we do know is that both Cruise and Waymo vehicles have a confounding habit of stopping, all the time, generally for unclear reasons and often in inconvenient places. The technology is, if nothing else, good enough to convince the CPUC board; one board member, John Reynolds, used to be a lawyer at Cruise. It is decidedly not good enough to keep the cars from stalling out and clogging the road whenever encountering an obstacle, which is to say "existing on a street, in San Francisco." AVs have driven into active construction sites and blocked traffic for hours, and the nearly 600 incidents logged by the city in the last 14 months are a certain underestimate.

Also, as the direct-action likers over at Safe Street Rebel discovered, you can shut these cars down with the use of the humble cone.

One would expect that transit and cycling advocate organizations like Safe Street Rebel would be against AVs, though what is most interesting about the coalition opposing their expansion is that it also includes many annoying city officials—remember, the vote is a state issue, so San Francisco officials have no power over the decision—as well as the fire department. Fire trucks famously prefer not to have cars in their way when they need to get places that are on fire, which makes constant t-posing obstacles like Cruise and Waymo vehicles a real hazard. "We’ve had them run over our fire hoses. We’ve had our hoses get caught in their axles. We’ve had them block fire engines, and we’ve had them come into live active fire scenes," San Francisco fire chief Jeanine Nicholson told the Washington Post. "We need something to change." In a public comment earlier this week, Nicholson also lamented how the SFFD was suddenly responsible for "babysitting" unresponsive AVs. Hilariously, the fire department took a cue from the activist set.

SFMTA chief Jeffrey Tumlim has also groused about the lack of data and busted incentives for the AV companies, while San Francisco Board of Supervisors president Aaron Peskin took it a step further. After the vote, he said that he and city officials would meet to discuss "next steps", the first of which he said would likely be filing for a rehearing, the first step towards legal action against the CPUC vote. "This is going to be an issue that San Francisco and cities and states around the country are going to grapple with for a long time to come," he said. "So this is the beginning, not the end."

As the Atlantic's David Zipper wrote, the safety issue is unsettled, though it is hardly the most relevant angle to the takeover of San Francisco by AVs. Will self-driving cars make it simpler to take car rides around the city? is a simple question with a simple answer; even an anonymous Safe Street Rebel rep answered it in the affirmative to the Atlantic. The harder and more relevant question is Do you want to be the sort of city where this stuff is ubiquitous? The particulars of the debate over the efficacy of AVs blows by the larger point that a new fleet of self-driving electric taxis still means more cars and more car-centric infrastructure. It means fewer bikes on the road, less support for public transit, and more pedestrian-hostile spaces—more of the same, but with fewer people involved.

When the privately owned automobile first hit American streets a century ago, it killed at an astonishing rate. Rather than rein in the technology, civil agencies captured by the auto industry instead took the ludicrous step of molding the shape of public infrastructure towards cars. As our erstwhile colleague Alex Pareene wrote a few years back, this is essentially the only path forward for the mass adoption of self-driving cars. They don't work, not really, under the current transportation paradigm; they can't, unless you can get rid of, uh, blossoms, road flares, and soap bubbles. Their only hope is not to get smarter but for their corporate sponsors to shift the paradigm in their favor.

That Waymo-ification does not mean safer streets, it means more cars, forever. It means more people siloed off, isolated in their bespoke consumer transit experience, apart from everyone else. What is the point of living in a city just to wall yourself off? The messy joy of urban life is its unpredictability, its communal nature. At its most elemental level, the car severs that. The only difference with onset of AVs is that they sever those connections more cleanly.