Yeah, I don't know why it has regressed to be so timid on these turns now. It's a full 10 km/h slower than it was when I tested it last spring. I know the new rewrite coming at the end of the year is supposed to be much better at labeling objects and handling more complex traffic situations, but it will be interesting to see if it's more confident on these winding roads as well.
Thanks for sharing your experience, @Durwood101 . I haven't encountered a situation yet where the light changes within that split second where you have to decide if you can make the light or need to slam on the brakes, but so far it's handled all of the early yellows brilliantly. One of the nice things about Teslas, though, is that you can easily make it through most yellow lights with that instant power if you want to.
@Silicon Desert : The jerky off-ramp issue has been happening to me for the past few software iterations, but I just did my first drive on 2020.36.3.1, and the two exits I made with NoA went much smoother than usual, so perhaps this issue has been ironed out. I'll have to test it on the usual exits to be sure though. Yeah, the normal highway stuff works nice and smooth for the most part, but I tend to push AP to its limits to see if it's handling the more challenging situations (i.e. merging onto the highway) better with each new software update. There definitely have been some serious regressions in the past - I think it was last year when a few iterations of the AP software were so jerky with the acceleration and deceleration in normal driving that it was unusable to me and very uncomfortable for my passengers. Luckily, that hasn't happened in quite a while.
I'm pretty sure the HW3/FSD computer has been stated to use around 25% more power than AP2/AP2.5. It sounds like the extra power drain that people are experiencing is due to how the car enters (or rather doesn't enter) sleep mode after the update. Oddly, I seem to be an anomaly since I'm finding the opposite situation where my car goes into sleep all the time now, and experiences less than usual phantom drain than before the update. Frankly, I'd prefer the phantom drain over waiting up to 1.5 minutes for my app to wake the car, but I'm going to get Tesla to take a look at this the next time I'm in.
Really? In my 4 years of having the car and using the app, I have never seen it take such a long time to connect. The longest I wait when the car is asleep is probably 15 seconds. Are you sure there isn't another factor?
It seems to vary quite a bit. I just woke the car now and it only took 7 seconds, but usually it takes 20, 30 or up to 90 seconds (maybe it's getting better?). Unfortunately, Tesla didn't have time to look into it yesterday since they squeezed me in for a next-day appointment to fix my horn and were very busy. Maybe I'll wait until the next software update and will book a mobile service if it's still happening after that.
I agree. If it's going to hug a line on a single lane highway, I'd prefer it be closer to the right side (shoulder). It gets freaky when it starts to move toward the centre line with oncoming traffic. I'm sure this will get ironed out in future software updates and I have high hopes for this major rewrite targeted in 6-10 weeks. I hope they're working on improving the lane keeping and making room for oncoming traffic in addition to the object labeling and FSD functionality.
Update: The car still seems to go to sleep very quickly and often after the MCU2 upgrade, but it now averages only 7-8 seconds to wake it via the app, which is totally acceptable, especially if it helps with phantom drain.
