Mary Cunningham, CBS News:

Tesla was found partly liable in a wrongful death case involving the electric vehicle company’s Autopilot system, with a jury awarding the plaintiffs $200 million in punitive damages plus additional money in compensatory damages.

[…]

“What we ultimately learned from that augmented video is that the vehicle 100% knew that it was about to run off the roadway, through a stop sign, through a blinking red light, through a parked car and through a pedestrian, yet did nothing other than shut itself off when the crash was unavoidable,” said Adam Boumel, one of the plaintiffs’ attorneys.

I continue to believe holding manufacturers legally responsible is the correct outcome for failures of autonomous driving technology. Corporations, unlike people, cannot go to jail; the closest thing we have to accountability is punitive damages.

⌥ Permalink

Jonathan M. Gitlin, Ars Technica:

Last night, after a wait of roughly an hour after the official start time, Elon Musk spoke to a crowd of Tesla fans and some journalists on a film studio backlot in California to give us an update on the company’s much-talked-about pivot to robotics. […]

[…]

After promising that “unsupervised FSD” is coming to all of Tesla’s five models — “now’s not the time for nuance,” Musk told a fan — he showed off a driverless minibus and then a horde of humanoid robots, which apparently leverage the same technology that Tesla says will be ready for autonomous driving with no supervision. These robots — “your own personal R2-D2,” he said — will apparently cost less than “$30,000” “long-term,” Musk claimed, adding that these would be the biggest product of all time, as all 8 billion people on earth would want one, then two, he predicted.

These announcements are almost certainly bullshit, and correctly contextualized by Gitlin. Mix the axiom “what can be asserted without evidence can also be dismissed without evidence” with the boy who cried “wolf!”, and the result is this media event — and that is without factoring in the usual Tesla sloppiness. These are three brand new products, all of which are purportedly future-defining, rambled about in the span of about thirty minutes on a random Thursday in October. Nothing is finished. Musk called two of the products “Cybercab” and “Optimus Robots”, but the company’s website refers to them as “Robotaxi” and “Tesla Bot”. Everything is hypothetical until proven otherwise.

The robot is particularly galling. The automotive industry has a long history of building humanoid robots: Honda’s ASIMO, Toyota’s Partner series, and General Motors’ work on NASA’s Robonaut 2. Some of these perform more specialized tasks. All of them have been around for a while. None of them are in widespread use. Tesla’s should be treated as an elaborate fiction until anyone outside the company can confirm even the most fundamental qualities it is claimed to possess.

Oh, and speaking of claims on the website, I want to address this:

To create a sustainable future, we must democratize transportation. We do this by making driving more efficient, affordable and safe. Autonomy makes this future possible, today.

Musk — for the featherweight of his words — said the Robotaxi would cost “less than $30,000” and be available “before 2027” — that is, to be clear, not “today”. If this thing ever ships, it will still require car-like infrastructure and ample space, even though it carries only two people.

Public transit, which is available today, is the very definition of democratized transportation, especially if it has been carefully considered for the needs of people with disabilities. It is inexpensive for end users, requires less space per person than any car, and has a beneficial feedback loop of safety and usage. I am not arguing the two cannot coexist; perhaps some of this stuff makes sense in low-density sprawl. But I have little confidence the future will look like Musk’s vision, or that Tesla will be delivering it. Why would anyone still believe this too-rich carnival barker who lies all the time?

⌥ Permalink