bhakan
MSLR MSM/Blk/19/FSD, Ord 6/10/21, Del 2/20/22
Waiting and sleepless in MarylandI don’t have the update and I’m one of the OGs from back when we had to play the Safety Score game. (Still have PTSD from it)
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FSD v12.x (end to end AI)
- Thread starter Buckminster
- Start date
-
- Tags
- #fixphantombraking autopilot chill is not chill elon is an ass fix fsd extreme late deceleration fsd does not commit to go fsd does not maintain proper speed fsd hard and late braking fsd jerky driving fsd late deceleration fsd stops too early too quickly fsd stops too far behind line fsd tailgating in city streets
j0shm1lls
Member
Shower thought: what if Tesla hates android users like me and pushes new builds to cars that are only attached to phones on the latest version of the phone app? (iOS is always ahead of android, currently iOS is v4.31 and android is v4.30.6)
You might want to consider a bath next time
yerEVan
Active Member
“Banana Republic?”
I’m personally keeping an eye on the pending value for 12.3 on TeslaFi to see approximately when the wave drops. It’s around ~30-50 from what I understand.
Hope it’s not just HW4 in the next one.
Hope it’s not just HW4 in the next one.
Last edited:
Hope this clears up some confusion about the highway stack in these FSD V12 builds and why people feel like their highway experiences are improved despite being on the old stack. The V11 highway stack still receives updates.
Still high on v12 but did have a BAD fail. Was on 14th street headed east and supposed to go straight across Peachtree St. but was incorrectly in the left turn lane. It even had 2 cars in front of it and both of course turned. It tried to go straight across, which is not really possible because the traffic going straight was backed up and there is a traffic island blocking. So had to disengage. Odd that I thought that NHTSA made Tesla honor wrong turns when in the turn lane. Even so there was no where to go except to turn.turned
j0shm1lls
Member
with or without toaster?You might want to consider a bath next time
Toasted water
i think i accidentaly put the fridge in the sink instead of the fork in the microwave
i think i accidentaly put the fridge in the sink instead of the fork in the microwave
I had an issue in Atlanta yesterday, the road forked and it went straight for the middle island...the path kept changing and I slammed on brakes about 2 feet from the concrete. The car just got confused.
For the past 2 years, my iPhone has never notified me when there's an update waiting, for some reason I always have to wake up my car and see if it's already on the screen
Surely this another sign of the pending apocalypse…9-mile commute. Link provided yesterday here which is important for evaluating the scenario. Pretty awesome overall. Light traffic.
0 disengagements (a first, by far, fewest ever is like 3-4 or something probably). There were one or two lucky situations where it was drifting uncertainly for a second, and if there had been cars around I would have disengaged.
~7+ interventions, for speed and a couple lane changes I wanted to make, or cancel impending incorrect ones it wanted to do.
Absolutely ZERO nags. Not one. (9 and 3 FTW.)
This thing is good (for my application).
To me, more investigation needed, but it seems to make some attempts to drift around the lane to create buffers, and it may be avoiding blind spots. BUT, I am not sure about this; obviously could be chance. Wasn’t super busy this morning so not too many opportunities to observe. TBD
Will it make mistakes? Yes
Is the speed control broken? Yes
Is the stopping behavior stupid? Yes
Will it be pretty useless in some scenarios? Yes
Will it be a robotaxi? Absolutely not on current hardware, obviously.
So it appears fundamentally good for certain specific scenarios, but is also currently broken, as previously described. We will see if it can overcome these issues and be useful in slightly more complex scenarios.
Yes, both modes are broken.
I had exactly this what yesterday, day before. Car was supposed to go STRAIGHT, but was mostly in the LEFT TURN LANE and just proceeded through the intersection from the LEFT TURN LANE WHICH HAD A RED LEFT LIGHT, and continued straight - which was the nav path.
j0shm1lls
Member
My post was satire more than anything.
Agree. After another hour today I found some hiccups (which I was expecting). Mostly related to somewhat blind unprotected / completely uncontrolled turns. Overall behavior is leaps and bounds ahead of anything prior for me. I tested most of my rough spots and probably like 90+% are fixed (for now at least).
Handled dedicated lanes for a turn with 'keep moving' signs where prior versions stopped. Slowed very slightly compared to what I'd do - probably an A-
Lane decisions drastically improved - though still probably a C+ or B- level
Stop signs continue to be amazing for me so far. Solid A
Couple roundabouts handled well without stopping. More evidence needed as these could be two lane but are generally just a single stream of cars in a wide lane. Rolled through very slightly slower than I would have. B+
If I understand correctly v12 has 3 parts:
1. City street stack
2. V11 enhanced highway stack
3. The part that handles the transition between city stack and highway stack. I think this part may cause early or late exit or incorrect move from ramp to highway.
sleepydoc
Well-Known Member
You’re not a member of the Junk crew by any means but I don’t think I’ve ever seen you this positive!9-mile commute. Link provided yesterday here which is important for evaluating the scenario. Pretty awesome overall. Light traffic.
0 disengagements (a first, by far, fewest ever is like 3-4 or something probably). There were one or two lucky situations where it was drifting uncertainly for a second, and if there had been cars around I would have disengaged.
~7+ interventions, for speed and a couple lane changes I wanted to make, or cancel impending incorrect ones it wanted to do.
Absolutely ZERO nags. Not one. (9 and 3 FTW.)
This thing is good (for my application).
To me, more investigation needed, but it seems to make some attempts to drift around the lane to create buffers, and it may be avoiding blind spots. BUT, I am not sure about this; obviously could be chance. Wasn’t super busy this morning so not too many opportunities to observe. TBD
Will it make mistakes? Yes
Is the speed control broken? Yes
Is the stopping behavior stupid? Yes
Will it be pretty useless in some scenarios? Yes
Will it be a robotaxi? Absolutely not on current hardware, obviously.
So it appears fundamentally good for certain specific scenarios, but is also currently broken, as previously described. We will see if it can overcome these issues and be useful in slightly more complex scenarios.
Yes, both modes are broken.
The good news is things like speed control and stopping behavior should be imminently fixable.
I’m curious to see how it behaves when extended to highways. For me one of the biggest limitations of V11 is the inability to plan exits. We have a lot of cloverleafs around here. Cloverleafs are admittedly bad design to begin with but on top of that we have several that require you to dart over as soon as the solid white line ends. V11 seems to take the approach that it can’t even start thinking about merging until the dotted line starts, then it needs about a block to build up courage before it gets enough nerve to try it and by that time you’ve missed the exit.
Similar threads
- Article
