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.
It'll be over 12 weeks since 10.12 was originally released, so presumably there's been 6 internal 2-week sprints that each could have been deployed if strictly following the release plan. Would you have been okay with 10.69 being called 10.18 instead to represent that there has been significant work more than "just" a regular release?

Or are you suggesting that Autopilot team isn't actually making progress in those 2-week sprints, so that's why there hasn't been releases?
Neither you nor I know what has been happening over in Fremont since 10.12 was released, and I see no evidence that there is a "2-week sprint" system. What I am suggesting that the progress since 10.5 has been extremely slow, and most of the problems that I hit on my daily commute that make FSD beta unusable for me and "nowhere near ready for L3" seem not to even have been addressed.
 
Better plug that thing in before you go road tripping.
The truck or the car? Probably you mean the former, 'cause we wouldn't want a phantom truck following us to the nearest super charger...

;-)

Seriously though, I've been letting the car get down below 20% and above 90% a couple times so that the BMS has a better idea of how much the battery can hold. Daily charges from 75 to 85% don't seem to give it enough slope in the curve to estimate where the top is. But that conversation belongs on a different thread.
 
  • Like
Reactions: FSDtester#1
How do you know they are at 100%. I hope it’s not because Elon said so lol
Because Elon said they were nearly there and it sounded pretty clearly that it was the goal. (No definition of 100% provided.) I’ve assumed therefore that their goal is effectively 100% (which of course would have to be simulated and could not be tested).
 
  • Funny
Reactions: KArnold
I had theorized that I had not been upgraded yet because I was opted into the beta, and that would have pushed us that were opted in past the branch point
That might match up with some official communication from Tesla replying to a request for FSD Beta back in February when 10.10.2 (2021.44.30.15) was the generally available version FSD Beta version:

Your VIN… is currently running a software version that is newer than the latest FSD Beta software version (2022.4 vs 2021.44.30.15) and we avoid downgrading software on production vehicles. The next FSD Beta release is currently planned for next week (V10.11) and will be newer than your current version. Therefore you could expect to download FSD Beta sometime in early March after v10.11 releases and assuming the SafetyScore on that car is still 97+.​
Best thing to do between now and then would be to avoid installing new software on the vehicle to prevent any unintentional leap-frogging.​
Although that last line about avoiding installing new software kinda sounds like even if Tesla tries to avoid sending out "too new" software for those who have pressed the request button, it could still happen and cause someone otherwise eligible for FSD Beta getting skipped. Then again, if many waiting in the queue are updating to 2022.20.8, then it might be safer to believe FSD Beta 10.69 will be based on that, e.g., 2022.20.8.5.
 
Because Elon said they were nearly there and it sounded pretty clearly that it was the goal. (No definition of 100% provided.) I’ve assumed therefore that their goal is effectively 100% (which of course would have to be simulated and could not be tested).
Maybe you’re being a bit sarcastic here, or a little loose with numbers—hard to tell. But be careful here. Elon never promised this upcoming version would 100% solve Chuck’s turn.

I know that at the shareholder meeting he said “and we’re going to solve Chuck’s turn, 100%.” But he never said it would be completely solved in this version as I recall. And by “100%”, I understood by context that he meant “definitely.” I think he would be the first to argue that FSD will never be 100% at anything. (Though he often speaks of being at least 10x human safety, which would imply a reliability in the very high 99.9x% range.)
 
no evidence that there is a "2-week sprint" system
FSD Beta updates used to be pretty consistently every 2 weeks although not necessarily always to the wider beta audience and definitely not so much since 10.11:
fsd beta updates.png


progress since 10.5 has been extremely slow
The improvements with 10.5 were indeed quite significant and noticeable for us as FSD Beta would no longer drive towards oncoming traffic lanes at some oddly shaped intersections. That version started using static world auto-labeling which was probably one of the first uses of leveraging all the data coming back since public beta 10.2. The auto-labeler with the existing neural network architecture was able to make significant improvements, but it probably started revealing cases that the architecture wasn't quite understanding easily for reasonable vector outputs resulting in the long development time for a new architecture coming to wider release with 10.69.
 
FSD Beta updates used to be pretty consistently every 2 weeks although not necessarily always to the wider beta audience and definitely not so much since 10.11:
View attachment 841557


The improvements with 10.5 were indeed quite significant and noticeable for us as FSD Beta would no longer drive towards oncoming traffic lanes at some oddly shaped intersections. That version started using static world auto-labeling which was probably one of the first uses of leveraging all the data coming back since public beta 10.2. The auto-labeler with the existing neural network architecture was able to make significant improvements, but it probably started revealing cases that the architecture wasn't quite understanding easily for reasonable vector outputs resulting in the long development time for a new architecture coming to wider release with 10.69.

Yea that makes sense. 10.5 was one of the biggest improvements for me as well.
 
True, however, when they complete this turn safely ~100% of the time like Elon clearly said they are going to, without overfitting, it may as well be the last version 10 anyway! That’s going to mean impressive capability improvements. Won’t be anything left to do but quickly add single stack in version 11 (going to be 🔥🔥🔥) and then a little cleanup with one or two more iterations, and wide release of completed FSD by the end of the year as promised. Book that revenue!
yup, I agree with that. Good point. I hope it really happens. I guess my point was poorly made that any software can in theory have hundreds of sub versions if they really want to. I am reminded by my son of an update to different type of software (not Tesla) lately that when we looked at the version number, it was 17.62.5706 :oops: Kind of makes me wonder if they really had 5706 builds of version 17.62 :D
 
I would assume the 20th would only be for the alpha testers, not to the fleet
At least for the last few versions (10.10, 10.11, 10.12), there's ".0" then ".1" then ".2" that allow for early testing and adjustments as well as partial rollouts starting with employees then private beta then public beta fleet random sampling, e.g., 20% to 40% to 100%.

If there's still 1 employee vehicle on TeslaFi and 2 on Teslascope, none of them have received an update from 10.13 to 10.69 yet.
 
At least for the last few versions (10.10, 10.11, 10.12), there's ".0" then ".1" then ".2" that allow for early testing and adjustments as well as partial rollouts starting with employees then private beta then public beta fleet random sampling, e.g., 20% to 40% to 100%.

If there's still 1 employee vehicle on TeslaFi and 2 on Teslascope, none of them have received an update from 10.13 to 10.69 yet.
by the time it gets to us it'll be 10.69.420
 
Elon said it would have to be deserving to be 10.69, and said clearly that it would solve Chuck’s turn 100%. So I guess if we get 10.13 they didn’t get there! So exciting to find out which thread to post in. Big changes promised.
If they don't and make it with a 69 release it will probably not be 10.13 since that was on 22.16 and they would likely have added changes to the Beta code since it was released. So it would probably be 10.14 at least.
 

One more sample test.

This one is well done. Failure per the terms though - slightly more than 5-second gap was missed on far side, should have proceeded at ~46 seconds after the jeep (usual caveats about potentially not seeing people who changed lanes, but we’ll see that in the full video when we actually see results from within the car).

Barely managed to avoid failure on near side - smallest gap was slightly less than 5 seconds.

Looks very good. Hopefully can be consistent - it’s been interesting to see the lack of consistency in prior videos.
 
Last edited: