Karpathy explained this at CVPR this year when introducing Tesla Vision:it chose to ignore and refused to act on avoiding the obstacle
We have a stationary approach again. … We are just approaching this vehicle and hoping to stop. What you see in orange in the legacy stack is that it actually takes us quite a bit of time for us to start slowing, and basically what's happening here is that the radar is very trigger happy, and it sees all these false stationary objects everywhere.
Radar by itself doesn't know what actually is a stationary car and what isn't, so it's waiting for vision to associate with it. Vision if it's not held up to a high enough bar is noisy and contributes sort of error. The sensor fusion stack just picks it up too late. So again you could fix all that -- even though it's a very gross system with a lot of if statements and so on. Because the sensor fusion is complicated, the error modes for vision and radar are quite different.
But here when we just work with vision alone and we take out the radar, vision recognizes this object very early gives the correct depth and velocity. There's no issues so we actually get an initial slow down much earlier, and we really simplified the stack a lot.