Uber's self-driving tech could just be an 'autonomous top hat' (GM)
Business Insider/Danielle Muoio
General Motors Vice-President Mark Reuss sat down for an interview with Business Insider in which the company veteran discussed Donald Trump, the strong US auto market, and self-driving cars.
His thoughts on autonomous vehicles were especially interesting and provocative.
GM's position on autonomous tech is closely tied with its acquisition of self-driving startup Cruise Automation in 2016 and it's $500-million investment in Lyft.
Reuss, GM CEO Mary Barra, and GM President Dan Ammann have all stressed that the company envisions fully autonomous vehicles arriving first in a ride-hailing configuration in a safe and controlled urban environment.
Uber is obviously pushing harder than that, first with a dramatic rollout in Pittsburgh last year, then with a controversial follow-up in San Francisco that ran afoul of the California DMV's autonomous vehicle licensing rules and ultimately led to Uber pulling its test fleet of Volvo XC90s off the road and to Arizona.
In Pittsburgh, Uber showcased Ford Fusion sedans with a complicated technology rig strapped to the top. The Volvos in San Francisco are more representative of where Uber wants to go with integrating the vehicle and the tech.
But GM's entire self-driving proposition relies on fully and safely integrating autonomous technologies with the company's manufacturing expertise.
"If you're going to attempt to put an autonomous top hat on a gas car and run it around, there's plenty of people who can do that," Reuss said.
"When you get into actually integrating an electrical architecture that depends on [artificial intelligence], that has some serious computing power behind it, and that's safe — you do all that on a car that's coming and on an electrified vehicle."
He added, "I don't see why you'd ever go back and retrofit things onto and old electrical architecture that's not designed for it. I don't see how you do that safely or in a way that's really the best."
Business Insider/Danielle Muoio
Reuss in particular and GM have made this point repeatedly, as has the traditional auto industry. In a nutshell, the rush to get self-driving cars on the road can't come at the experimental expense of safety.
Integrating autonomous technology into a platform such as the Chevy Bolt electric vehicle, which started arriving at dealerships in California and Oregon late last year, solves the safety problem.
Reuss didn't argue that Uber's approach doesn't work — and in any case, he wasn't naming names when it came to Lyft's $60-billion-plus-valuation competitor — but he did insist that the top-hat approach isn't the one that GM will be taking.
"Yes they can run autonomous," he said. "But ultimately you need to have enough computing power on board to drive deep machine learning. Those are powerful tools that everybody can have, but that also requires a deep integration of systems in a car to make it safe."