Uber is giving up on self-driving cars-they just won’t admit it yet – Info Gadgets

Uber CEO Dara Khosrowshahi maintains that the company is committed to its autonomous-vehicle program — most recently while speaking with Kara Swisher at the Code Conference on May 31 — but he said the same about not leaving international markets after SoftBank’s investment in late 2017, then announcing a deal with Grab to offload its Southeast Asian operations only a few months later.

While Khosrowshahi won’t admit it, the company’s reaction to the fatal Arizona collision in March and its recent emphasis on expanding Uber’s focus beyond its ride-hailing business to include a broader suite of transportation offerings suggest it’s leaving behind the passion project of disgraced cofounder Travis Kalanick and embracing the alternative vision of its new management team — and that means the failing autonomous-vehicle project is on the backburner.

Big problems with Uber’s self-driving vehicles

As much as Uber wanted to be a major player in the race for self-driving cars, it’s been clear for quite some time that the company’s efforts weren’t anywhere near its competition — that created pressure to find “cheat codes” and develop a “strategy to take all the shortcuts we can,” as Kalanick discussed with Anthony Levandowski, the executive who led Waymo’s self-driving team before making the jump to Uber (and allegedly taking a bunch of documents with him).

Uber’s subpar autonomous system has long been a concern, to the point that the company was ranked among the worst in the industry by Navigant Research at the end of 2017. That designation was confirmed in the aftermath of the fatal collision on March 18 in Tempe, Arizona, when details were leaked which showed that Uber’s test vehicles were struggling to go a mere 13 miles (21 km) before a human had to take over, compared to the average of 5,600 miles (9,000 km) per disengagement that Waymo reported at the end of 2017 for its operations in California.

The leak also showed that Uber’s Arizona self-driving team was under intense pressure to deliver a workable system for public launch by the end of 2018 and to provide a demonstration to Khosrowshahi at the end of April — neither of which they were prepared for. Even as the team knew their tech wasn’t up to par, they cut the number of safety drivers from one to two, meaning a single person had to monitor the road and the vehicle’s diagnostics.

Even more damning was the report released by the National Transportation Safety Board (NTSB) in May, where the agency described how the vehicle’s sensors detected the pedestrian six seconds before impact, but didn’t determine it should brake until 1.3 seconds before impact because the system was unable to classify her, thinking she was an unknown object, a bicycle, then a vehicle. But even when the system determined it should brake, it couldn’t do so; the Arizona team had disabled emergency braking in autonomous mode to ensure a smoother ride, in case the system tried to stop if it misidentified something. They also didn’t install an indicator for the system to alert the safety driver if she needed to brake.

After the crash, Uber halted its testing of self-driving vehicles. The company settled with the family of the pedestrian to avoid a court battle and a legal precedent that could have gone against self-driving vehicles. Soon after, Khosrowshahi announced Uber would permanently cease operations in Arizona — the governor would likely have opposed allowing their test vehicles on public roads again anyway. Khosrowshahi says testing will resume elsewhere in the coming months, but it’s hard to see how Uber has any chance of catching up to its competitors.

The leaks about the company’s technology make it clear that its system can only go very short distances before a human has to take over, and its progress may have even gone backward. Further, as part of the settlement in the Waymo-Uber lawsuit, Uber has to “ensure that Waymo confidential information is not being incorporated into Uber technology,” which could place an additional barrier on the development of its autonomous system.

While Uber may be unlikely to end its autonomous-vehicle program entirely, at least in the near future, it will likely scale down its ambitions in the space as it pivots to its new plan to be the dominant transportation platform with a whole host of options for its users, both first- and third-party. Khosrowshahi gave a hint as to what this might look like in his recent conversation with Kara Swisher.

Uber’s pivot to a new vision

Uber is a wildly unprofitable company — it lost $2.8 billion in 2016 and $4.5 billion in 2017 — and investors are pushing Khosrowshahi to find a path to profitability. Kalanick saw self-driving cars as that path: by replacing drivers with computers, the company would be able to lower fares even further, while still making it to black because it would essentially eliminate payouts to drivers. Khosrowshahi has a different approach.

Initially, he was not convinced by Uber’s focus on self-driving cars, but told reporters that changed after a series of conversations he had after assuming the role of CEO — whether that was his true conviction, or simply something a new CEO says after taking over a dumpster fire is unknown. However, it wouldn’t be surprising if the fatal collision in Arizona gave him the ammunition he needed to change course and push his own vision for Uber’s future.

When Swisher brought up Kalanick’s view that computers would replace Uber’s drivers, Khosrowshahi said, resolutely, “this is something I fundamentally disagree with,” before talking about the importance of human drivers — “[t]he face of Uber is the person sitting in the front seat” — and how he instead sees the technology as being able to “augment” humans on Uber’s “hybrid network.”

But the real change in strategy became obvious when Khosrowshahi described autonomous-vehicle technology and its place on Uber’s platform. He asserted that Uber’s autonomous-vehicle testing will continue in future and said the company is “not going to look to own the technology for ourselves and will license it to third parties.” When Swisher pushed back, asking if Uber “[had] to make” the technology, Khosrowshahi responded that

we have to have access to it. And I think there are gonna be many autonomous players, and that’s why I think as a principle, we will license out our own technology, and then we’ll look to build around other autonomous technology as well. We’re neutral. We’re a network company. So, if GM builds autonomous technology, I’d welcome Waymo to put cars into our network as well. We wanna be totally neutral.

This is such a far cry from Kalanick’s view that Uber had to cut corners in order to “win” the self-driving race. Maybe that’s because Khosrowshahi knows there’s no way for Uber to win by developing its own technology — it’s way too far behind — so, instead, he wants to make sure autonomous-vehicle technology is open and licensed, not controlled by a single company. It remains to be seen if the leaders, notably Waymo and GM, will agree.

Khosrowshahi’s stated desire to have the autonomous vehicles of other companies on the Uber app is indicative of how his vision differs from that of Kalanick. Under new leadership, Uber is aggressively expanding its transportation offerings — some of which it controls, but others that it doesn’t.

The Amazon for transportation

Expansion is nothing new for Uber. Its core service has expanded over the years from a black car service, to a taxi replacement service, and most recently to offering shared rides on semi-fixed routes — Khosrowshahi has even mused about Uber branching out into buses, which the company tried in Egypt. Uber drivers can participate in food and package delivery, and other transportation modes — such as motorcycles, rickshaws, helicopters, and ferries — have been added in some international jurisdictions.

These last examples were simply a recognition of how transportation varies in different parts of the world, while Uber’s new vision of becoming the dominant platform for the future of transportation is something else entirely. Instead of adding services focused on just a few cities, its new plan is for services that it can roll out relatively quickly across a large portion of its network.

In April, Uber bought Jump, a dockless bike-share service operating in San Francisco and Washington, DC. Khosrowshahi told Swisher that the average length of a Jump trip is 2.6 miles (4.2 km), while 30 to 40 percent of Uber trips in San Francisco are that length or less, so Jump could begin to cannibalize some of those shorter Uber rides. Uber is already planning a European rollout for Jump, and applied for a permit to offer a scooter service in San Francisco — yet another mode to add to the app.

Uber also added public-transit ticketing in some cities through a partnership with Masabi, which Khosrowshahi says he hopes to expand even further — though it remains to be seen whether Uber will want a cut of every ticket sold through the app. And, while the future of Uber’s self-driving vehicles is unclear, Khosrowshahi is trying to get other companies’ services on the app as they launch. He also added car sharing for users in San Francisco through a partnership with Getaround.

This doesn’t mean Khosrowshahi doesn’t have his own big bet. Instead of focusing on self-driving vehicles, the company is now placing more attention on its flying-car project, Uber Elevate, even though it’s highly unlikely to deliver the equity or be available on the timelines Khosrowshahi is promising. The Skyports the company showed at its recent conference are also incredibly car-centric and fail to integrate into their surrounding urban environments.

However, that’s just one mode in a much larger vision that Khosrowshahi has begun to execute, and if cities and transportation advocates were concerned about what Uber was doing to urban transportation with its ride-hailing service, this new approach could bring many more issues.

Let’s be clear: making it easier for people to access more options for active transportation — bikes and scooters — is a great move, but where’s the evidence that Uber is working with local governments to ensure these services are rolled out in a way that enhances city or county transportation plans?

Further, Uber already has a lot of valuable ridership data — and will soon have much more — that could help cities more effectively plan for transportation within their boundaries, but even though Khosrowshahi has promised to work more closely with local governments, there’s been little action on Uber’s part to share its data or by governments to compel it.

There’s also a bigger question that needs to be asked: if transportation is going digital and being integrated into a single platform, who should control it? A private company or the local transportation authority? This isn’t just a question about Uber, but a more fundamental one about how we approach digital platforms: should they be owned by private companies that can give themselves an advantage, or should they be publicly owned, making it easier to ensure neutrality and serve the public good? No wonder there’s a renewed conversation about antitrust and what role large corporations should play in our societies — it’s one we desperately need to have.