What are peoples thoughts on whether this gets released to the entire fleet?
My initial impression was that it would never get released to the entire fleet because it seems primarily geared towards data collection. But, I couldn't think of a case where a passive method wouldn't have worked just as well. I decided to stop thinking about it, and to come back a few days later to see if anyone answered that. But, I don't see any good answers so far that justifies the existence of stopping at every light for training purposes.
So now I'm wondering if I was wrong, and that it will get released after validation. It fulfills the "Recognize and respond to traffic lights and stop signs." requirement of FSD so that means Tesla can likely recognize additional revenue from FSD. Even though it's not what we were expecting.
Is that really a bad thing? Is it any worse than what we have now?
What we have now will blow through a light or a stop sign if the user isn't paying attention. Sure sometimes it will warn them, and the warning might or might not give them enough time to slow down. So in a sense switching to this would be better. It might be annoying enough that those who use AP on roads with stop lights eventually get annoyed enough that they stop doing that.
I simply can't see using this on a road with lots of traffic lights.
Now where I do see this useful is on highways with occasional traffic lights. Even if the car does start to slow gradually before a green light I don't think it's necessarily a bad thing. If the person in front of me slowed a bit before a green light on a high speed road it wouldn't really bug me. I'd likely think "Oh, nervous are we?"
It might cause a few rear end crashes, but I think overall this method will reduce accidents over the entire fleet.