[pdk42]

I've spent a good part of the morning reading through this topic from the beginning. And so far it seems that anyone who has 8.0 then gets a chance to experience 8.5 doesn't like it. So even though BMW says you can't do it, or won't do it, doesn't seem to matter because the difference is minimal, and whatever advantages it has over 8.0 the majority who've experienced both prefer the older version. Most of the argument seems to follow the philosophy of why the won't make the upgrade possible, not that it's "in demand" because of all the new features it has, at least this is what it seems like after almost 3 years of opening this discussion. A lot of interesting takes and complaints, but I just got my 23' a couple of months ago and I love the car, and since I have nothing to compare it to, the version 8.0 works fine with my needs, I can't think of any scenarios where any updates to menus or displays would change my mind, but you never know what the future will have to offer. Maybe one day when we're all gone they'll be saying "remember when they used to put screens in cars"?

I quite like some of the 8.0 UI elements and concept even though my car has 8.5. I guess the only area where 8.5 is better in my view is having the main menu icons always visible at the bottom of the screen. But I don’t think 8.5 is really any better overall. It wouldn’t bother me if I had 8.0. Also, 8.0 can be coded more extensively by Bimmercode etc, so in fact, reflecting on it now, I think I’d prefer 8.0!

As far as FSD goes with Tesla or Waymo, I think it's a joke to think this is going to work with the complexity of everything involved not just within the technology to make it work. From my layman understanding from what I've read and watched experts talk about FSD, they all agreed that Tesla's camera only system will never work safely enough to be fully autonomous and without the use of LIDAR. Look at the Tesla Robotaxi experiments going on right now, they are not doing as well as Waymo and those aren't reliable either. I personally think FSD is like a marketing scam, just to try and say "we did first" kind of thing.

I think that the use of neural nets is the right approach to driving automation and that we will eventually get there. But the vision-only approach of Tesla’s is really never going to work properly. There are too many limitations and environmental situations where it’s going to have insufficiently clear images to do its job.