Wednesday, April 24, 2019

Elon Musk's self-confidence shouldn't inspire much confidence

More good work from Timothy B. Lee on the Tesla beat [emphasis added]:
Tesla is less than two years away from full self-driving, CEO Elon Musk said in an interview with MIT researcher Lex Fridman published on Friday. And he said Tesla was far ahead of other companies working on self-driving technology.

"To me right now, this seems 'game, set, and match,'" Musk said. "I could be wrong, but it appears to be the case that Tesla is vastly ahead of everyone."

Musk told Fridman that Tesla customers would need to keep their hands on the wheel "for at least six months or something like that." But he predicted that soon—"maybe even toward the end of this year, I'd be shocked if it's not next year at the latest"—Tesla's self-driving technology will become so good that "having a human intervene will decrease safety."

Musk has maintained an optimistic mood about Tesla's self-driving progress at the same time that other industry CEOs have been tamping down expectations. (Musk is congenitally optimistic on this topic—in 2015, he predicted that fully self-driving cars would be ready within two years, declaring it a "much easier problem than people think it is.")


"Tesla Autopilot is not yet even close to where Waymo was 6 years ago," wrote Brad Templeton, a longtime self-driving car evangelist who advised Google during the early years of its self-driving car program.

By 2012, Google had developed highway driver assistance software that had capabilities similar to recent versions of Autopilot. Google considered selling this technology as a standalone product but decided there was too much danger of drivers becoming overly reliant on the technology—and failing to properly monitor it.

So Google pivoted to developing fully self-driving cars that would never need customers to take control. By 2015, Waymo was confident enough in its technology to allow a blind man to ride through residential streets in Austin, Texas. But more than three years later, Google's project—now rechristened Waymo—still hasn't launched a fully driverless taxi service.

At this point, regular viewers should be able to fill in the the missing dialogue without much trouble. Musk is a huckster, and like most of the best of the breed, he has the gift of believing his own line. His natural lack of talent as an engineer (he really is shockingly bad at this) may actually give him an advantage here -- that Dunning–Kruger boost of confidence can make a pitch even more convincing.

You can almost imagine the conversation with the genuinely gifted engineers at Tesla as they tried to explain the many problems with that were almost solved compared to the handful that were still years away from resolution, patiently walking him through the subtle distinctions and spelling out the difficult trade-offs required to get an autonomous vehicle ready for the road.

The real tip-off is "much easier problem than people think it is."  If you've ever had a job that required presenting complex technical or analytic concepts, you've probably had that moment where you realize that your audience missed all of those fine but important point and drew exactly the wrong conclusions.

Of course, your audience probably didn't use those conclusions to justify the viability of a multi-billion dollar company.

No comments:

Post a Comment