As FSD's capabilities improve, it is more important than ever to intervene or disengage whenever the car doesn't operate as you'd expect a human to do the same maneuver. There is no stronger signal we can send than disengaging.
This is going to require some serious effort on Tesla's part...
Many of us disengage FSDb
prior to entering areas that it struggles with. Before the report button was removed, I'd let FSDb struggle through the area in question a time or three, hit the report button, then from that point on disengage FSDb before entering the area it struggled with/deserved a "report."
I don't have that option anymore. So now when I find an area that FSDb struggles with, I just disengage prior to entering that area. These disengagements are far more numerous than the initial disengagement when the area of poor performance was found, and vary in the timing of them a lot; sometimes I'll disengage quite a ways before the troublesome area.
A lot of guys on the forum have stated that they do the same... disengaging FSDb before entering a situation that it struggles with.
So now they're not only going to have to look at the situation that existed prior to the disengagement, they're also going to have to look
ahead quite a bit. Is that even possible? Can they look at the upcoming driving situation and estimate what FSDb would have struggled with to the point that the driver didn't want to use it in that scenario?
Wouldn't this be more burdensome than the heaps of data they have to sift through to separate the wheat from the chaff from all the "reports" they get?
Have they even considered that people disengage prior to areas it struggles with?