I actually agree with you. But most people with enhanced autopilot have HW2.5, and now with all the resources on HW3, I expect EAP will always be a crap shoot beta. I would love to have FSD NoA for the freeway, and park assist and summon that actually work, but I don’t think that’ll ever happen for HW2.5 owners
if FSD reaches a point where subscriptions models are a thing than I imagine they'll use that to get HW2.5 people over to HW3. Something like 6 months free FSD with a purchase of an upgrade. So even if they don't continue on with FSD after those 6 months they fall back to the EAP feature set. An EAP feature that's not the old code, but the new code the requires HW3.
I bought well before the subscription model was mention by Elon. So I have EAP+FSD because I felt like paying $3K was justifiable for the promised upgrade to HW3. In fact out of the EAP+FSD the only thing of any real value that I feel like I've gotten so far is HW3.
As to actual autonomous driving I don't believe HW3 is capable of it due to weather limitations with 5 out of the 8 cameras. The 3 forward looking ones are good. The rear view one is the best I've ever had in a vehicle, but still not good enough. What concerns me the most are the side ones as they are the ones that always trigger the "camera blinded" error message. There also isn't any redundancy so I think regulators simply won't allow for truly L3 (or beyond) autonomous driving.
So it's always going to be L2 driving.
But, I think that's fine due to all the issues that need to get ironed out.
Most of why I'm unhappy with the current EAP feature set is for nav, and nav interpretation reasons.
With Smart Summons I was able to improve it at least for the parking lot at work by updating the map in OpenStreetmaps. Now I never managed 10/10 or even close to 10/10 on performance (the same start and same stop point), but it was better.
For NoA I'm trying to rile up the local WA state people to push Tesla to fix broken things. In fact I'm thinking of funding a prize winning challenge myself. I'm seriously thinking of giving (or a chartable contribution )$500 to the first person who drives on NoA (with unconfirmed lane changes with a set speed of at least +5mph) from Portland, OR to Everett, WA with zero interventions. I'll put a time limit on the challenge so that if no one wins the primary challenge they still have a chance of winning $500 for the closest attempt.
My feeling is the lack of feedback from customers to Tesla is why NoA sucks so bad in places outside of California.