It looks like you're more of a glass-half-empty person (and me a glass-half-full). I agree with your current assessment of how good or bad fsd beta 10.69.x.x is. However, as I see it, the rate of improvement is such that FSD will reach L3 autonomy in 2 years (and requiring one hardware update).

Especially because perception, which I guessed would be the hardest part to solve due to lack of LIDAR, seems to be pretty good already, it's the planning that's bad.