Anyone who is still thinking to trust Tesla AP/FSD should watch this video [1]
Most of the Tesla's high stock valuation depends on them executing Robotaxi biz plan and seeing how Waymo cars are self-driving in some cities for so long and now beginning on highways, I still wonder why is taking them so long to understand the proven first-principle that LIDAR and Radars are essential hardware components to solve self driving and they just can't rely on cameras alone.
I get the sneaky feeling the idea of fsd with cameras is just an idea that is WILDLY ahead of its time. The technology isn't there yet, but will be eventually.
It's the middle part that's scary (where we are now I guess)
If camera based self driving doesn't happen at first, it never will later.
That's because it would be a safety downgrade, and the public never accepts new technology which does the same as a previous one, yet drops a safety feature.
For example, requiring a handbrake separate from the main brakes in cars. There hasn't been an engineering reason to keep these separate for 50 years, the main brakes already are doubly redundant. But people won't adjust laws to reduce safety, even if the safety reduction is so vanishingly small so as to not matter.
The issue with vision only, especially without IR is that even if it becomes safer than a human, it won't ever have the awareness of vision+IR/Lidar, thus be less safe.
Most of the Tesla's high stock valuation depends on them executing Robotaxi biz plan and seeing how Waymo cars are self-driving in some cities for so long and now beginning on highways, I still wonder why is taking them so long to understand the proven first-principle that LIDAR and Radars are essential hardware components to solve self driving and they just can't rely on cameras alone.
[1] https://youtu.be/V2u3dcH2VGM?si=iUvrnZWgYK9stylv