I took delivery of a 2023 Y long range at the end of January. I put in my order perhaps a week after they dropped the price $15,000 in Canada and took delivery about only 10 days later. Checking the claimed durations would give you a bead on my mental sharpness LOL. I like the car. I bought it with Enhanced Autopilot and am toying with spending more to get FSD which I realize is a misnomer at this point. The worst disappointment to me so far is the software update that gave me park assist as part of autopilot. Despite the disclosure before purchase that I would not immediately be getting automatic parking or the automatic summon functions I had still expected a car with something like the feature i had to wait for weeks to get by the early April software update. I did a test drive of a new Model Y with only 400 km on it being sold by a Hyundai dealer (they had 2 - and must have got them through deception) shortly before my purchase decision. I offered the Hyundai dealer the dropped price by Tesla and the Hyundai dealer wanted the old price from some poor clueless soul they could fool. The interesting thing is that the car I test drove had better park assist than what I got from the early April download for my new car. The old car must have had USS. Anyhow, the new car has no USS and the quality of the park assist at this point is worse than useless for getting in and out of my tight garage. Maybe they can do better with future software. The idea I gather is to learn by experience how to assess distances from different simultaneous camera images over time as the car moves. So initially I'd hoped the software might markedly improve over time, but it hasn't and I can officially grade it to be worse than useless for getting in and out of my garage. It makes a racket of complaint as I back out repeatedly telling me to stop. I'd never get out if I paid attention. When you think about it, USS is all at the front of the car, so the problem backing out has to be pure software deficiency. The rear view side cameras are great so I don't see a need for such a problem. On the highway I find the enhanced autopilot useful on freeways. I use it on city streets as well, subject to having to help it with intersections and sometimes other things. I have not done a deeply systematic assessment but I like having it. I'm interested in buying FSD so I can have a beta version which I know is far short of true FSD but which I think I'd still like having. Foolishness? I need to do more evaluation. I tend to believe that with the hardware I have I'll never have fully functional FSD given that HW4 which is coming will support a number of more cameras, and these will each have something like 4 times the resolution of my cameras. Beyond that the new computer in HW4 has a form factor that makes fitting it into the same space as my present computer unlikely. I've watched some FSD videos on Youtube (not enough) and I wonder at the detail in the depiction of traffic - which exceeds the detail of the depictions of traffic on my screen. That may just be software but I wonder. The view of vehicles to the left and right of intersections interest me. That would have to be based on the cameras on the side of my car between the front and back doors. The location of those cameras seems less than great being so far back, but when you think about it, they are only slightly behind the head of the driver who traditionally was the one making traffic judgements. My sense is that the software I am now using just does not use those cameras in driving and if I bought FSD I'm guessing those cameras would come into force for the first time for building the traffic images on my screen. I saw some guy labouring through the thought process of how the new cameras coming could not possibly in his mind work with the same driving software as the old cameras now in place, and I think his thinking is flawed. I have a good background in software and I do not think there is a sound basis for his concern that he will be left with nothing. To make software maintainable it is built in modules that perform very particular functions quite distinct from other modules and functions. It makes testing far easier as you just separately determine that each piece does its job. The job of the camera system seems to me to allow the building of a dynamic model of the reality of the road and the traffic within the computer.. Other quite distinct software would do the job of figuring out what the car must do from this model of the world in response to the driver's instructions and the objectives being served by the self driving system. The separation of these functions in software should mean that you can upgrade the cameras and the software for building a model of the world the car is in, without a total redo of the software that tells the car features what they each must do in response to that world to meet required objectives. So, yes, I'd expect there to be ongoing limitations from my HW3, but I'd also expect some inferior level of ongoing functionality that would not be lost as more advanced hardware and software were put in place. That's my theory, and for now I'm sticking to it. LOL Finally, once the FSD objectives are finally largely met at some yet remote time - the price may get worse. It could turn a car into a taxi without a labour cost. My car is unlikely to be good enough for that without different hardware, according to my best guess. They are not upgrading hardware because they don't have to and just want to make a show.
