Yes and possibly. If Tesla can get vision to provide the same performance as USS on existing cars then they will turn it off.
That's the trouble with this casual kind of discussion. Small changes make big differences.
I don't see them maintaining two stacks though. One for new cameras-equipped cars, and another for 'legacy' USS...
I am quite active in the German speaking forums. I have personally been told by two independent agents at the Tesla sales center that they expect USS to be shut off when the TV solution comes. There are many reports from others around Switzerland, Austria and Germany that say the same. We also believe they will not maintain 2 stacks. For the 2023 cars now being delivered without USS, we see still the “old” cameras. There are no deliveries yet with new cameras. We believe that when USS is shut off in favor of the “to be delivered in 2 weeks Elon time” TV solution, those with old cameras will have a “good enough” park assistance function, but probably nowhere near as capable as the cars with new cameras. There is also speculation that LiDAR is quietly being installed on the 2023 builds, but that to me is still wishful thinking.
I'm cheating, but here's my trip into Reading this weekend, 166wh/mile
www.teslarati.com
Given current discussions, it is ambiguous, but I think the article is still talking about camera upgrades to the very first type. No mention of 5mp. I doubt they will be made available on current cars..... although if needed for FSD I think there is a good case that they should be.Tesla offering Autopilot Camera Upgrades to owners with FSD capability
Tesla is offering Autopilot camera upgrades to FSD customers with high Safety Scores. The upgrades come just before FSD's wide release.
Here is yet another proof point that the TV solution will rely on new cameras. New cameras that currently are not installed on new 2023 cars being delivered at least in Germany without USS. If you have not taken delivery yet but are about to, maybe re-think that. Seems like new cameras are critical for FSD, so may also be critical to reinstate a park assistance function that is at parity to what USS offers.
It is certain that the AI will be trained without the USS going forwards, thus use will be removed there. For the distance visualisation shown on screen, it is quite probable that that will continue to use USS. From a software perspective, the vision solution can just simulate the existing sensor information (i.e. the distance to objects) that will be rendered. Then in the same software release, you just need an optional software flag to have that data provided by the existing USS code or the new vision based calculations - so there will be no need for 2 stacks as such.
The occupancy network algorithms are more efficient than current vision, so there may not be a need for a better CPU.
It is shocking that they still think of Tesla as a startup. That may be the core of the issue.
Hmm. Too much reference to 'bloat' there and the sensors role in AI. These sensors provide value to the consumer. Unless you provide a suitable alternative to the consumer for an expected feature in a 50k + car then you have failed in your role as an engineer.
How far ahead do you look when driving? Your brain is capable of picking up a driving related anomaly and focussing on it in a way that essentially enhances detail. You would need 200mm telephoto or equvelent hi Res to equal that..in stereo. Just think about the judgements you make on country roads when planning to overtake a tractor...assessing distances to curves, rate if acceleration, other road hazards such as blind field gates....
