Before the announcement a few of us predicted the demise of EAP.
I was one of them, but I didn't think they'd actually do it. I thought it was something they needed to do, but I didn't know if the financials worked for it.
The reason for this was Tesla didn't seem to have two code bases. They just had a much more advanced neural net that didn't work in HW2.5. At least not with frame rate, and complexity that it should have.
We also saw Tesla hitting the limitations of AP2 where certain features like Dash Cam, and Sentry mode simply don't work. So I think it was evident that they were hitting the limits.
So Tesla had to move over to HW3, and still grow features within EAP without upsetting current EAP/HW2+ owners.
How do you do that? You kill EAP all together for new vehicles, and you simply deliver a final build of it (not counting simple maintenance releases). Where that final build takes care of the remaining features. (like unconfirmed lane changes).
Then you give people with EAP some incentives to get them to upgrade to FSD. Tesla did that by killing EAP, and charging $2K for the update from EAP to FSD.
Of course it's not really a full upgrade from EAP to FSD as they moved features from EAP to FSD. So that $2K isn't what I'd call a discount, but a fair price for additional features. It's also a great price since it should come with a HW3 upgrade.
From a technical standpoint I think it makes a lot of sense. It's still a bad name, but at least they have room to grow.