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

FSD Beta 10.69

This site may earn commission on affiliate links.
I figured out the release date, decided to cancel monthly beta testing fee which expires on Sept 12 - that will be the mega rollout of 10.69.
Soo this indeed happened, I only had 1 drive the morning of Sept 12 with 10.69.2 which I already talked about. Got the message on the screen about software config / subscription change so I thought it was over .... but it turns out I am stuck in FSD beta mode (for free) and I can't turn it off! Hah. On the autopilot setting screen none of the FSD beta options are available - noting to toggle on or off. Tried to switch to my non-beta profile - no luck there either.
 
10.69.2 really isn't that much better than 10.12 IMO.

It's much much better in turns and unprotected turns.

But overall, there's too many unnecessary brakes and a couple of annoying aspects that make it not comfortable / fun / safe to use compared to 10.12.

Just one example: merging behavior onto crowded highways = just tries to rush to the front, avoiding all other cars like they're obstacles, lol. Even when there's no merging lane left, it's still trying to push past the front car.

I also enjoyed the 10.12.2's general highway performance more as well.
 
10.69.2 really isn't that much better than 10.12 IMO.

It's much much better in turns and unprotected turns.

But overall, there's too many unnecessary brakes and a couple of annoying aspects that make it not comfortable / fun / safe to use compared to 10.12.

Just one example: merging behavior onto crowded highways = just tries to rush to the front, avoiding all other cars like they're obstacles, lol. Even when there's no merging lane left, it's still trying to push past the front car.

I also enjoyed the 10.12.2's general highway performance more as well.
not mindblown anymore?? Oh well, look forward to your comments on the next version.
 
I think you misunderstood my post - I said that pressing the accelerator was essentially as bad as a disengagement and in response to a follow up post said that it didn’t matter whether it was because the system was slow or just unable to proceed. I also hypothesized that pressing the accelerator *may* send a report to Tesla.
I'm not sure what you mean by "bad." You original post seemed to be discouraging accelerator use. Anyway it seems like accelerator use without disengagement is the most easily utilized information that Tesla can receive as you're letting them know exactly where the error was and what control output is necessary to resolve it.
 
  • Like
Reactions: Silicon Desert
not mindblown anymore?? Oh well, look forward to your comments on the next version.

A couple of "minor" issues can make a version annoying to use. My two most annoying issues:

1) accel and decel in certain situations is too jarring / slightly nauseating
2) unnecessary braking in random spots, difficult to predict or anticipate
 
I'm not sure what you mean by "bad." You original post seemed to be discouraging accelerator use. Anyway it seems like accelerator use without disengagement is the most easily utilized information that Tesla can receive as you're letting them know exactly where the error was and what control output is necessary to resolve it.
My original post was simply meant to point out that 'just' pressing the accelerator is actually a much more significant intervention than most people realize since you're doing all the decision making that FSD is supposed to be doing.
 
Had a weird situation with 10.69.2 build. We have a side by side garage, that's about 100ft down the driveway. I tend to back our cars into the garage, and last night when taking the Model 3 out for a drive the Lane Departure Alarm went off as I was driving forward out of my garage at <5mph. Then as I straightened out the wheel to drive down the driveway, it went off again.

Hadn't seen that on 10.12.2 nor 10.69.1.1, so I'll try to see if it happens again tonight.
 
  • Funny
Reactions: EVNow
Something that 69.2 still doesn't do well (or at all) is multiple rapid lane changes. I've had a couple cases in the last few days where I turned right, then had to get over 2 lanes to make a quick left. FSD completely fails in these cases.
lane change 1.jpg
Lane Change 2.jpg
 
Something that 69.2 still doesn't do well (or at all) is multiple rapid lane changes. I've had a couple cases in the last few days where I turned right, then had to get over 2 lanes to make a quick left. FSD completely fails in these cases. View attachment 852262View attachment 852263
I have a similar situation with a right hand turn onto a three-lane frontage road where beta needs to get in the left lane to enter a tollway. 10.12.2 was the first beta version to turn into the center lane and usually made the change to the left lane with little problem. 10.69.2 is back to old behavior turning into the right lane, moving quickly to the middle lane, then slowing to a near stop despite there still being space to make the final lane change.

Maybe 10.70 will be better.
 
  • Informative
Reactions: sleepydoc
Drove a lot today, on a profile that doesn't have FSDb active, but checked my router tonight and the car dumped a TON of data to the mothership. Wonder what triggers I hit. Mostly a drive into the city and back. A bit rainy.

pure speculation, but I wonder if this is "shadow mode" data to help with single-stack release.
My wife used the car yesterday afternoon, same difference in profiles. Ours did not send anything back after the drive. I had the same question when I saw that, because I was under the interpretation that it was running in shadow mode when FSD is not enabled.
 
Why are you talking about Lidar? We're talking about HD MAPS

Because the original post was talking about lidar to generate HD maps

Post that prompted this tangent said:
Apple uses LIDAR generated HD maps

And I was pointing out LIDAR wouldn't actually help with any of the lane definition issues that is the majority of recent complaints.

LIDAR is a crutch for someone who can't figure out getting vision perception right.

Once you have and correctly ID lanes (as it appears most folks are saying Tesla has) it's the planner is the issue- and LIDAR doesn't help with that.
 
in order to not to get distracted by parked vehicle problems
Ha, I just encountered an unnecessary lane change with 10.69.2 not too dissimilar to Kim's curving left except there were no cars nearby, so no issues with lead vehicles. It was a simple curve left where I was in the right-most lane out of 4 lanes total. A message popped up for "path blockage," but there was nothing in the road. Pretty sure it was the occupancy network mispredicting the trees ahead as being in the current lane and there wasn't enough visibility to see that the trees were "obviously" past the curb, which was just out of view due to the slight cresting of the road:

path blockage.jpg


This is from a Model 3, so maybe the extra 7.1" height difference for a Model Y would have avoided the unnecessary lane change in this case.
 
Pretty sure it was the occupancy network mispredicting the trees ahead as being in the current lane and there wasn't enough visibility to see that the trees were "obviously" past the curb, which was just out of view due to the slight cresting of the road

With all the false slowdowns with 10.69.2, I can see Tesla taking a long time to widely deploy single stack. If they deploy it before end of year, there'll likely be a lot of phantom braking complaints lol
 
  • Funny
Reactions: AlanSubie4Life
Took it on a short drive. Didn't notice anything too different, its maybe 1% smoother. Still made the same mistakes it usually does for me.

Steering sensor sensitivity doesn't seem any different to me and my wheel weight still works btw.

Can't believe we waited so long for this. I think a lot of the YouTubers have a placebo effect on them. You could give them 10.1 and call it 10.69.420 and they will all rave about how smooth and amazing it is compared to before.
 
Took it on a short drive. Didn't notice anything too different, its maybe 1% smoother. Still made the same mistakes it usually does for me.

Steering sensor sensitivity doesn't seem any different to me and my wheel weight still works btw.

Can't believe we waited so long for this. I think a lot of the YouTubers have a placebo effect on them. You could give them 10.1 and call it 10.69.420 and they will all rave about how smooth and amazing it is compared to before.
Shills will do anything to get the hits and responses from EM
 
Steering sensor sensitivity doesn't seem any different to me
Yeah I tend to agree, after another day of driving it. Basically seems to be just fine if you hold the wheel. 10.69.1.1 was a whole other level of torque requirement though.

Hopefully they’ll defeat your wheel weight soon; should be easy.

——
Post-69 refractory thoughts:

In retrospect we* should have known 69 with Elon would be dissatisfying. **

* I always knew and was consistent in my conviction. May even get two beers, but not counting my chickens.

** .69 is definitely better than .12 in terms of turning, and positioning on unmarked wide streets, as I said a week ago or so. Everything else seems pretty broken. Simply can’t make some trivial turns. (Can’t get into double left turn lanes, N=2.)
 
  • Like
Reactions: FSDtester#1