Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Firmware 8.5

This site may earn commission on affiliate links.
Tested on my morning commute with auto lane-change enabled (no turn signal confirmation). I have my setting at 'mild' fwiw.

I noticed the biggest improvement in when AP now does lane changes. On one stretch of road, 5.15 would sometimes suggest a change to "Follow route" that, if taken, would have put me in a lane that became Exit Only before my desired exit. It didn't happen once this morning. It also didn't suggest a lane change to a faster lane to the left as I got near one exit; under 5.15 it almost always suggested a change. This would then have made it very difficult to get back over to the exit.

I did have to abort one lane change to the right as an 18-wheeler signaled that it was going to change to its left into the same lane. The truck had only signaled and wasn't moving over yet but I didn't want to wait and see how that movie ended.

Overall this seems like a significant step forward but I'm probably going to go back to turn signal confirmation for now.
 
It does seem smoother, even coming from 8.4 (which I only had for about two days). I'm glad that I can set AoN to be the default now. I probably won't turn on lane change without confirmation, I feel like AoN wants me to swerve back and forth between lanes too much (especially when I'm clearly overtaking someone and it wants me to get out of the passing lane behind them) and I like to be able to just ignore its suggestions when I think it is appropriate.

Edited to add:

Well, that probably explains why it feels smoother: Elon tweeted yesterday that 8.5 actually has a neural net upgrade. That's kind of a big deal.
 
Received 8.5 last night. Tested NAP with auto lane changing this morning.... Had to take over 75% of the lane change attempts. :( This has been a problem for me since 2 updates ago. The vehicle signals a change, attempts to change, and then panics and moves back into the original lane. Does this a couple times before I just finally have to take over so other drives don't think I'm hammered. There was nobody in the lane each time this happened. Cameras are clean, sensors are clean. What gives!?!?!
 
Received 8.5 last night. Tested NAP with auto lane changing this morning.... Had to take over 75% of the lane change attempts. :( This has been a problem for me since 2 updates ago. The vehicle signals a change, attempts to change, and then panics and moves back into the original lane. Does this a couple times before I just finally have to take over so other drives don't think I'm hammered. There was nobody in the lane each time this happened. Cameras are clean, sensors are clean. What gives!?!?!

I still have a theory about this, so let's see if your situation fits it: Where was the sun in relation to your car/the lane you were change into?
 
Did about 60 miles round trip in NOA. It worked very well with only one disengagement in fairly light/medium traffic. There was one weird section on the 91E where I was in the left lane adjacent to the carpool lane and NOA kept prompting me to change lanes to the right and then immediately cancelling the prompt before I could confirm and did that over and over and over again for no discernible reason.

I’d also swear the regen is slightly more powerful, too. I’m pretty sure I’ve never seen the regen line go as far as it was today, and it seems like the car is slowing better too.
 
  • Informative
Reactions: jrad6515
I’d also add that the vehicles displayed around the 3 in the UI seem more stable. I passsed a few semis and didn’t notice nearly as much of the image splitting in two and hopping around all over the place as I did before when passing them. Looking forward to testing it out on my way up to SF next weekend.
 
My theory (base solely on my experience with this situation) is that the sun can blind a camera or two with lens flaring. Nothing more than speculation based on anecdotal evidence though.

My experience is very limited, since I've only been driving my Model 3 since Tuesday, and so far this has all been on 2019.8.4, with AP but not yet NoAP. I've had to intervene in about 1/2 of the car's lane changes; it either doesn't begin a change in time (when I'm trying to get into an exit lane) or begins a change but then aborts it for no apparent reason. Many of these instances have been at night, so there's no possibility of the sun causing lens flares; however, there's always the chance that other cars' headlights are doing the same thing. In either case, if that's what's going on, it's a problem that Tesla must overcome if the AP/NoAP features are to move beyond the "toy" phase. As it is, I'm reluctant to use this feature at all; the aborted lane changes must be quite annoying to other drivers, and could actually create a hazard in heavy traffic. That said, I'll keep trying it for a while longer, especially once I get 2019.8.5 -- and beyond.
 
  • Informative
Reactions: hcdavis3
I got 8.5 last night. Car didn't make any lane changes on its own this morning while on NoA, so I guess I have that turned off. But, when I try to use the stalk to change lanes, it still does the hesitation dance. Been that way since it went to 15.5
 
Aborted lane changes first appeared for me in 2019.8.3 and continue in 2019.8.5. Prior to that lane changes had been amazingly competent and NOA has been great. I'm certainly concerned that lane changes are now broken. I concur with above posts that it seems to be having difficulty in reading the lane markings of the new lane for some reason.
 
I didn't think a reset would mess with your display brightness settings?

Someone commented to Sheryl Crow's tweet that the reboot should have "fixed" it and Elon Musk replied that it was odd that it didn't.

I don't know the answer but the bottom line is Elon said they would update the FW so that there were never situations in which the screen would be unusable. The car can use the ambient light sensor for example to force a minimal visible screen brightness during the day, regardless of what some idiot user thinks they ought to do (in this case the idiot user was apparently Crow's child).
 
  • Like
Reactions: Mdoss