my guess is that the models are simply doing what humans do: glare in your eyes you slow down. Of course the car can still see fine but the humans that it's trying to mimic can't ... so the FSD ends up slowing down too.

Interesting idea. Sounds a bit similar to what Musk has described as a difficulty for neural nets when they're trying to learn how to stop at a stop sign, given the fact that the humans that they are imitating usually don't:


Musk explained the challenge posed by the fact that less than 0.5% of drivers come to a complete stop at stop signs.

This makes it incredibly difficult to program the vehicle with such sparse data on people obeying the law. This lack of data highlights Tesla's obstacles in training its AI systems to operate in real-world scenarios where human drivers frequently do not adhere to the rules.


Stock down today with the flare-up of Musk vs Trump flaming. I think both of them will back off pretty quickly, and it is all a side-show. The only real question is do we get to FSD soon or not. And, maybe I should add, to what extent does Tesla have a significant advantage in robotic manufacturing and robots in general, which is hard to judge. I still think I want to be on the Musk side of the bet and I took advantage of the price swoon today to get up to 1.5% of my portfolio. Maybe I will double that if we get down to $250 a share, which we might, since I expect the news in the next few months to stay bad. But the real question is what happens in 12-24 months, and I think this is Tesla's game to win.

DTB