When the $70 billion tech giant Apple abandons an idea, it may be an idea simply not worth pursuing, which could be the state of the suddenly struggling autonomous vehicle industry.

It’s been an astonishing and quick fall from grace for self-driving cars, in the six months since the California Public Utilities Commission allowed them to have unlimited expansion in San Francisco. Since then, one of the GM-owned Cruise autonomous cars dragged a pedestrian 20 feet, which has effectively sent the company into freefall, as GM is losing an estimated $2 billion a year.

Google’s Waymo has also had some stumbles, and Tesla’s full self-driving mode has repeatedly shown itself to be error-prone.

And news broke last week that tech giant Apple was shutting down their electric-car department, a secretive effort called “Project Titan” which was also a self-driving electric car — because the company decided a while back to try to create their own superior version of a Tesla with autopilot. Apple hasn’t said much about why they shut the whole thing down. But when a cash-flush company like Apple, with a $3 trillion valuation and hundreds of billions to spend gives up an an idea, maybe the idea just isn't that lucrative in the long run?

That’s the focus of a just-published New York Magazine piece entitled “The Self-Driving Car Bubble Has Popped.”

“Tesla has blown through countless promises by Elon Musk that fully autonomous cars were just one more year away and is facing fresh scrutiny after an engineer died while using his Model 3’s ‘full self-driving” feature,” as author John Herrman writes. “In October, Musk admitted that he’d been ‘overly optimistic’ about the technology. Uber stopped developing its own AVs. Ford and Volkswagen abandoned a joint project, Argo, after billions of dollars of investment.”

And remember, Uber also claimed back in 2016 that they would have flying cars by the year 2020, a claim that is laughable in retrospect.  

The autonomous vehicle technology seems to have plateaued, getting things right about 90% of the time, but unable to iterate on that difficult last 10% (the “last mile,” to use an industry term). And as Herrman notes in New York Mag, the most useful AI driving features are those that assist human drivers, like “adaptive cruise control, lane assist, emergency braking.”

In other words, don’t look now, but the humans might actually win the self-driving car wars.

Related: Cruise Recalling Self-Driving Robotaxis Nationwide, Amidst Report Vehicles Had Safety Issues Around Children [SFNews]

Image: @SafeStreetRebel via Twitter