No. of Recommendations: 3
I have the guy on mute so I don't know everything that was said -- what's with Tesla investors and their need to make this a competition? -- but since it was quote posted, I stand by my point:
Tesla is still years away from making good on their promises of full self driving, as indicated by the plethora of other videos out there showing, in a relatively easy driving environment, and with barely any miles earned, a robotaxi (1) going into the wrong lane after bailing on a left turn (a nice honk from another driver to emphasize the infraction), (2) ceasing service due to rain (kind of a big deal), (3) possibly clipping another car while parking so the supervisor has to get in and take over, (4) a CLEAR intervention needed by the so-called "safety monitor" to avoid an accident with a UPS truck(5) and braking dangerously for no reason (sun glare) or because there are emergency vehicles way off on the roadside. We're days in with barely any vehicles -- that's a lot of risk piling up. We'll ignore that the only people who got to get inside were TSLA investors and superfans, so who knows what happened that we didn't see or hear about.
What is Musk, "funding secured", claiming? It is slightly ambiguous. My guess is that they had a remote supervisor able to take over if the car needed it but they didn't need to intervene. Maybe they didn't have a supervisor at all (if so, this kind of stunt is why Texas just passed a law, to join basically every other jurisdiction that matters). Regardless, the event is unremarkable if you've been paying attention -- Teslas (and many other cars!) have been doing stretches of self-driving for a while. I know the exact road travelled and I'm confident Tesla so-called FSD can usually do it without intervention. It's a long way from there to being able to do what a robotaxi needs to do, from a safety and regulatory perspective. The data we have -- much of it crowdsourced because Tesla won't share it for some mysterious reason -- shows they aren't close to being able to hit the "mileage without intervention marker" that is necessary to run these things safely without supervision, and in limited use cases.
Maybe they can eventually do it camera-only, maybe they can't. My engineering friends are more than skeptical.
In sum, all of this is years away from camera-only autonomous driving that can be hailed or used on demand, rain or shine, sun glare or no glare, without killing people at a decent clip. The video evidence from the good-weather only rollout suggests as much. I'm sure Musk will promise otherwise. He also said these things would be on the road in 2016, so I don't exactly hang on his every word when looking for the truth of things.
