AP1 was just the MobileEye platform transplanted into MS/MX. It had years and years of development and success prior to that. Also, EAP and FSD wasn't promised for AP1, so 11 months after AP1 was released, it pretty much met most of the expectations set when it was released. They were called "Autopilot Convenience Features". AP2 OTOH, has been sold and promised to deliver full EAP and FSD by now, and it hasn't even reached full AP1 parity yet. And no way is even the Great and Powerful Elon Musk going to be able to summon his car, fully autonomously, from LA to a parking lot in NYC by the end of this year. And if he does, it's going to have so many disclaimers and exceptions, they might as well not call it FSD.
Quick disclaimer, I realize what Tesla has promised and what was delivered is entirely two different things. However, You need to turn that around and rethink what it looks like from side of ap1 operation.
Mobile eye (ap1) was in development for about a decade before in 2014 Tesla said they will put their hardware into their car.
In 2015, ap1 was finally enabled.
What was ap1 stability in 2016, 11 months after it was enabled, didn't they actually have someone die? I realize it was actually two traffic deaths (1 in China, 1 in Florida).
Here comes Tesla in 2016 after deaths and says, let's replace Mobile eye and delivers their own hardware.
What is ap2 stability in 2017? About the same as ap1. Here is why:
in stark terms, no one died. Let's hope it stays this way.
In gray terms, ap2 stability on road at 11 months is about the same where ap1 stability was 11 months after it was enabled.
Why do we think Tesla engineers are capable?
Tesla Ap2 team did not have a decade to get their expertise like mobile eye team and yet stability is close. Parity is somewhat close as well.
You tell me 80/20 rule, yes that's a good rule until you realize that mobile ap1 80% is somewhat stable lvl2, while Tesla ap2+ 80% is lvl 3+. If you have ap2+ Tesla, im betting by 3-5 year of ownership, lvl3 will be available + some lvl4.