This spring, Uber and Lyft are ushering in a new wave of IPOs from gig-economy companies that are reorganizing the physical world as well as the digital one. Over the next several weeks, The Ringer will be examining the impact these companies have on the way we travel, live, and define the concept of a job in the modern age.
For Elon Musk, the driverless car is always right around the corner. At an investor day event last month focused on Tesla’s autonomous driving technology, the CEO predicted that his company would have a million cars on the road next year with self-driving hardware “at a reliability level that we would consider that no one needs to pay attention.” That means Level 5 autonomy, per the Society of Automotive Engineers, or a vehicle that can travel on any road at any time without human intervention. It’s a level of technological advancement I once compared to the Batmobile .
Musk has made these kinds of claims before. In 2015 he predicted that Teslas would have “ complete autonomy ” by 2017 and a regulatory green light a year later. In 2016 he said that a Tesla would be able to drive itself from Los Angeles to New York by 2017, a feat that still hasn’t happened. In 2017 he said people would be able to safely sleep in their fully autonomous Teslas in about two years. The future is now, but napping in the driver’s seat of a moving vehicle remains extremely dangerous. “At best, they may be able to create a system that functions under certain limited scenarios. It will not be fully autonomous in 2020 or anytime in the next several years.” —Sam Abuelsamid, research analyst for Navigant In the past, Musk’s bold predictions have been met with A-for-effort enthusiasm and a smattering of polite skepticism. But the response this time has been different. People have less patience for PR campaigns masquerading as engineering timelines. “That’s bullshit,” says Sam Abuelsamid, a research analyst for Navigant, a consulting firm that ranks companies on the viability of their autonomous vehicle plans. “At best, they may be able to create a system that functions under certain limited scenarios. It will not be fully autonomous in 2020 or anytime in the next several years.”
What’s changed? Self-driving cars—and their associated building blocks such as machine learning, computer vision, and LIDAR —continue to improve, but executives other than Musk have been admitting that reports of their impending deployment were greatly exaggerated. Ford CEO Jim Hackett said last month that the industry had “overestimated the arrival of autonomous vehicles.” Chris Urmson, the former leader of Google’s self-driving car project, once hoped that his son wouldn’t need a driver’s license because driverless cars would be so plentiful by 2020. Now the CEO of the self-driving startup Aurora, Urmson says that driverless cars will be slowly integrated onto our roads “over the next 30 to 50 years .” That’s nearly as long as it took computers to evolve from IBM’s first mainframe to Apple’s first iPhone.
Despite the setbacks, money continues to pour into autonomous vehicle research. Companies from several industries involved have a lot to gain as long as the self-driving dream remains just beyond reach. For the tech giants, the driverless car could be the moonshot that finally pays off after a decade-plus focused on consolidation rather than innovation. For Detroit automakers, the technology is a hedge against the disruptors in Silicon Valley. For Uber and Lyft, it’s the silver bullet that will turn money-losing businesses into profitable ones at an unspecified future date. This kind of rosy prognostication has led to breathless headlines and massive influxes of investment dollars. Ford just nabbed a $1.7 billion investment in driverless cars from Volkswagen. Uber’s driverless car division was recently valued at $7.25 billion . And Waymo, Alphabet’s self-driving development subsidiary, is seeking outside investment that may value the company at “ several times ” General Motors’ autonomous division, which is worth $15 billion.
Yet the precise business models these companies will use on driverless cars remain murky. What’s more, this lack of clarity comes partially by design; for now, there’s more money to be made on the idea of driverless cars than on the vehicles themselves. “I don’t think it’s all PR, but I do think it’s a bit of wishful thinking,” says Mike Ramsey, an analyst at Gartner who specializes in autonomous vehicles. “When you have platforms that have been disruptive in their use and changing the dynamics of how people consume mobility, but no clear path to profitability, there’s a desire to look for technology to help solve that.”
All of these companies, from Detroit to Silicon Valley, are pitching a similar driverless utopia: Instead of owning cars, people will be whisked through city streets by autonomous vehicles that can operate 24 hours a day, reducing vehicular fatalities and traffic congestion in the process. When Google originally introduced the driverless car as a practical research effort in 2010, the company projected that it would be able to save as many as 600,000 lives per year and double the capacity of roads. And though Tesla wants to sell autonomous vehicles to individual consumers, most companies have determined that assembling an army of centrally controlled robotic taxis makes more financial sense. A corporate fleet could be constrained to carefully mapped-out areas, and the expensive cameras and sensors necessary to operate the cars could be paid for with recurrent revenue from shared rides.
This utopia doesn’t exist, however, because there are a number of technical and legal hurdles that autonomous vehicles still have yet to surmount. The basic challenge of allowing driverless cars to be able to accurately perceive the world around them has not been solved; GM’s vehicles continue to have trouble distinguishing moving objects from stationary ones, according to an October 2018 report from Reuters . Beyond that, the AI brains of these cars have to be able to make split-second decisions based on a constantly changing environment. Given the risks involved, engineers have programmed some driverless cars to be overly cautious , stalling out behind unforeseen obstacles such as construction sites, or braking jarringly like a teenager in driver’s ed. Finally, Congress’s failure to pass a federal law regulating autonomous vehicles late last year makes widespread deployment of the technology impractical, even if the kinks get worked out. “You have more grounded expectations about the timeline for these technologies, combined with a lack of industry push for legislation,” says Bryant Walker Smith, a law professor at the University of South Carolina who studies legislation of autonomous vehicles. “A lot of companies don’t yet know what they want.”
For Uber and Lyft in particular, the driverless car is easier to manage as a just-on-the horizon technological myth rather than an actual product. Both companies are deeply unprofitable and have long propped up autonomous vehicles as the next key step to both revolutionizing mobility and achieving riches. Back in August 2016, former Uber CEO Travis Kalanick said that the implementation of autonomous vehicles on the platform was “ a years thing, not a decade thing .” But Uber’s $680 million acquisition of the self-driving truck startup Otto embroiled the company in a vicious legal battle with Google over allegations of stolen LIDAR technology . In March 2018, one of Uber’s driverless cars struck and killed a pedestrian in Tempe, Arizona, prompting the state to immediately ban Uber’s test vehicles. Investigations into the incident by The New York Times and Business Insider revealed that Uber’s autonomous vehicles at the time required human intervention more often than their target of every 13 miles and had their ability to suddenly brake disabled to provide for a smoother riding experience. Prioritizing safety over ambitious launch targets might have prevented the fatal collision.
Uber, which said the vehicle’s brakes were disabled to prevent erratic driving , was charged with no crime in the Arizona crash and reached an undisclosed settlement with the victim’s family. The company’s AI leaders, since chastened, will now only say that deploying driverless cars at scale will take “ a long time .” “There’s pressure when you have an exciting technology to set a stake in the ground and say we’ll be done by this point,” Ramsey says. “When you get in reality and you have to make a decision about unleashing a technology where literally if it fails people die, the standard for being done is a little bit elusive.”
Lyft’s foray into self-driving has not been nearly as disastrous, but the company has similarly pared down expectations. In 2016 Lyft president John Zimmer predicted that the majority of Lyft rides would be autonomous by 2021. So far it’s served only 40,000 rides in driverless cars as part of an ongoing Las Vegas trial, a total that accounts for a mere fraction of the more than half a billion rides it provides annually. Though the company just announced a partnership with Waymo that will bring a limited number of driverless cars to the Lyft platform in Phoenix, those rides will still have a human driver sitting as a fail-safe in the front seat, just as there are during the Las Vegas trial. In its March IPO filing, Lyft put its projection for majority-autonomous rides at 2029.
In addition to the shifting timelines, there also remain pressing questions about what driverless cars would accomplish, economically and socially, if or when they are unleashed on the world. Autonomous vehicles would theoretically allow Uber and Lyft to cut costs by eliminating drivers and rapidly scaling up their operations. But reaping the full benefits of autonomous vehicles would require the ride-hailing giants to own and maintain their own vehicles, a massive new expense that would amount to a totally new business model.
The alternate option, which Lyft is pursuing with its Waymo partnership, is to serve as a platform for other companies that operate the cars. But that could end up simply replacing the expense of a human driver with the cost of a robotic one. Aurora, the AI startup billing itself as the brains of the driverless vehicle, plans to get paid on a per-mile basis, just like human drivers do today. Companies “are going to want to be compensated for buying and operating the autonomous vehicles,” Ramsey says. “So I really don’t see how autonomous vehicles actually change the business model of what they’re doing that much.” Lyft’s Zimmer also said that the rise of driverless cars would “all but end” private vehicle ownership in major cities by 2025. But Lyft’s and Uber’s previous promises to help eliminate congestion in cities have proved false, according to a variety of studies . In a world where they and other firms have implemented autonomous ride-hailing fleets at scale, congestion could in fact increase, as individual ride-hailing customers would likely continue choosing single-occupant trips over shared rides, and the number of cars on the road could shoot up further if driverless food and package deliveries are cheaper than those executed by humans.
“That’s the dystopian view of what happens with automated cars,” Mark Hallenbeck, director of the Washington State Transportation Center at the University of Washington, told me when tech execs in Seattle were floating the idea of driverless cars replacing public transit . “It fits really well with the view of how humans interact with technology. We’re not interested in the social good of everybody working together. We’re interested in me.”
Most people probably won’t have to think about the societal effects of driverless cars for a while. Abuelsamid predicts that the average person won’t be able to summon a driverless car in the U.S. until the late 2020s at the earliest, and companies’ increasingly sketchy timelines seem to support that claim. That could provide time for government regulation to catch up to the impacts that Uber and Lyft are already having on cities before a slew of similar driverless services are unleashed. And the pressure to be safe seems to have eclipsed the pressure to be first, a […]