I've only been watching this space for a couple of months. It looks to me that they took a major pivot onto a new FSD implementation at some point and expect AP and EAP to migrate to that once ready.
They appear to have made the decision to put all their AP/EAP/FSD team behind the new stack and the AP/EAP we use daily is essentially dead in the water waiting for the new world to come along.
This decision wouldn't have burned so bad if the optimistic time table had been even approximately correct, but the never-ending push back of a shipping date makes the negligence of AP/EAP totally embarrassing.
I work in software too, and deprecated services become a major headache if the replacements are substandard or late.
Definitely not an excuse - the decision was a really bad one in hindsight, the new FSD had better be good and better be here soon!
Not totally sure there is any good reason Tesla could possible have for such a crap piece of software (or hardware) for their adaptive cruise control!!!!
Let's try to look at this from an engineer's viewpoint because that is what I am. I do both hardware and software engineering as well as mechanical. Like I said above Toyota had a good adaptive cruise control over a decade ago and what sensor did they use ? Radar. Same with 2014 Mitsubishi Phev I owned, radar, and no problems encountered like I am seeing on my model Y except for a very heavy rain event making the radar unusable. But that is totally understandable to me because as a pilot of high performance pressurised twins I have seen radar rain shadows and know to stay away from them. So I know radar as a sensor. I have seen it in action in multiple uses and car manufacturers have made it work as a successful sensor in all but extreme cases (heavy rain).
So why not Tesla? They used radar. Now here we might get into excuses here like maybe: But we want to have a total Full Self Driving package blah blah blah.
And this to me as an engineer and product developer is where Tesla are doing it wrong. I never use my customers as product beta testers. Do all your testing as much as possible in house. Hell I have spent massive amounts of time doing that with all-nighters and all weekends when schedules are tight doing tests. Sure you may still get caught out with some really unforeseen circumstance later that may require a mod but nothing like I am seeing here with Tesla.
I have seen their ex chief software engineer from memory going on about how all vision hardware will probably be the way to go to finally end up as a good FSD package by keeping it simpler blah blah blah neural networks blah blah blah.
Well they may be right in the long run but prove that in house. To me I want to say get back to basics Tesla. Get your cruise control sorted. Maybe some like me don't need all that other crap like FSD. Maybe some of your customers like me just don't want our wife, girlfriends, family or whatever having the crap scared out of them when you can't get your act together and get something basic working like a reasonably reliable cruise control.
Ok rant over but just had my oldest son tell me about 2 major 110kph to 0 stopping events on a trip to Sydney from just going under some overpasses from his description and this terrified his fiancé. May have been lighting from time of day blah blah blah but it's got to point where I need to think about banning the use of cruise control or put my family through a major training exercise in how I control these events as much as possible or possibly selling the Y.
But this shouldn't be needed in this day and age of such a reasonably old technology like adaptive cruise control.
And again, all this is just my opinion.