As a software guy, the fact that they are still making architectural changes in FSD leads me to believe that they are further away from releasing this than just fixing a few bugs here and there from the beta feedback. Some assumptions they originally made were wrong for one reason or another, and the underlying code needed to be changed. It could mean that they figured out a clever new way to do something, but either way, these are not the kind of changes you make if a wide release is imminent.
I fear you might be right.
I think there was a big rewrite around 2 years ago, going from still frames to video. That also took a while.
Oh well. This is a huge endeavor.
Elon is pushing as hard as possible - he is not holding back.
But, in a weird way, solving FSD in 2-4 years might actually be more lucrative than solving it in 6 months.
FSD is still, to many just another tech-nerd pipe dream. It might very well not work. Why bother?
When (
if)Tesla solves FSD they will instantly have a huge, fluorescent target on their back.
Tesla pre-FSD is just dangerous and annoying to big oil. Tesla with FSD: this is now a
mortal threat.
You think fudsters, big oil and even country actors are fighting Tesla hard now? Just wait for Tesla to solve FSD, and their campaign to delay or sabotage Tesla will reach a fever pitch.
The higher production capacity Tesla has when FSD is fully ready, the better.
That is why Elon is ramping factory capacity and battery production so hard: He cannot forecast the arrival of FSD with precision, but knows that whenever it does arrive Tesla will need to operate in plaid mode.
Of course, the huge downside to Tesla being 'late' on FSD is that competitors might come first. But, as has been discussed before, most of them, except comma.ai, are relying on lidar and maps, and that just takes longer to scale.