The object persistence would have to take account of all of the things we don't currently see on the visualisation; walls, trolleys, rocks, plants, etc. If it does start having to remember all of this stuff, just because they want to get rid of ultrasonics, is that a good thing? I see it as broadly the same argument in favour of the $5 Bosch rain sensor - how much development time and processing power is wasted trying to reinvent a working system?
Based on how "effective" auto high beams and auto wipers are, I'm not confident.
The best thing you can say about either of those two systems is that sometimes they reach parity with other car's systems using traditional sensors. Some people report few problems with them, others report regular problems, but there is definitely not the "it just works" consensus that you should (and can) expect with these systems in 2022. I suspect there is a narrow envelope where both systems on a Tesla work effectively, much narrower than the aforementioned Bosch rain sensor. I can't even really explain why the auto high beams system is so retrograde.
Yep it would, and I don’t even think the latest FSD is picking up really small objects that could still cause damage. Then again sometimes the sensors don’t.
I agree I think it’s a huge software overhead just to remove some sensors and it may not be a cost saving after all. I can’t fathom it either really, then again I’m not an AI megamind, but then again then again I’ve worked with people like that before and whilst they’re due their dues sometimes ones drive to best a challenge blinds us to real world practicalities and the question “why”.
Either way it’s a wild ride owning these things. I’d be much more annoyed if I was a customer paying for EAP or FSD. As it is I’m just not installing any more updates for a while.