FSD V11 Discussion - First Released 11/11/22 at 11:11PM - Maybe

(moderator note: Added maybe to the title. This may be an alpha or beta release, so “release” is relative. Stay tuned.)

We finally have confirmation of the release of FSD V11. This presumably is single stack; we’ll see, as the question was left unanswered. I guess it will still be called FSD Beta V11?

https://twitter.com/x/status/1591336830389067778

( )

 

[Twiglett]

On a total aside, I was reading an article on Aviation and Automation. Apparently, regulators are pushing for more automation to help reduce the load on pilots (and potentially eliminate co-pilots for some flights).

I recall many discussions about the term "autopilot" in Tesla vs the aviation industry. And this observation from a pilot was an eye opener, and drew parallels to Tesla and some of the issues AP/FSD have had:

from the last example - the automated landing system had a phantom braking event?
Interesting that even in the world of highly trained professionals that complacency takes over.

 

[JulienW]

While it is within the realm of possibilities that V11 could roll out to all FSD users before the end of the year here is a FAR more likely theoretical scenario timeline for such a LARGE update that will likely need lots of work.

No way would I count on V11 making a full wide releases as a holiday release (unless you count St Patrick's Day) .

• V11.0 release to influencers on Dec 16th
• Several bugs identified
• V11.1 released to influencers on Jan 6th
• V11.1 released to a few Beta users on Jan 13th
• Rollout stopped, likely more bugs
• V11.2 released on Jan 20th.
• Slow rollout over the next 2 weeks to Beta users
• V11.3 released on Feb 17th
• After a couple of weeks in March Tesla decides to allow it to go to all FSD users

 

[kriskros91]

Releasing without a safety score requirement (which sounds like it will happen “soon”) does not necessarily equal a wide release, as discussed.

Obviously the timeline could be off too, but that is neither here nor there.

I'm truly not trying to be a PITA but I don't understand the distinction.

If I am a new FSD purchaser ($15000 is a LOT though) or a new subscriber to FSD in January 2023 I expect that my car will IMMEDIATELY have a new toggle to use Full Self-Driving (beta) as it is in our cars after we went through the ordeal of Safety Score and endless waiting.

Are you saying something different to that?

**Edit to add January 2023 and not just "2023" which could obviously have a totally different meaning.

 

[kriskros91]

While it is within the realm of possibilities that V11 could roll out to all FSD users before the end of the year here is a FAR more likely theoretical scenario timeline for such a LARGE update that will likely need lots of work.

No way would I count on V11 making a full wide releases as a holiday release (unless you count St Patrick's Day) .

• V11.0 release to influencers on Dec 16th
• Several bugs identified
• V11.1 released to influencers on Jan 6th
• V11.1 released to a few Beta users on Jan 13th
• Rollout stopped, likely more bugs
• V11.2 released on Jan 20th.
• Slow rollout over the next 2 weeks to Beta users
• V11.3 released on Feb 17th
• After a couple of weeks in March Tesla decides to allow it to go to all FSD users

This is possible. What I think is more of Tesla's style is that it will actually be narrow release to influencers starting this week (before Thanksgiving as Elon stated). Then throughout early to mid December there will be wide(r) release to those with FSDbeta already and those that have been doing Safey Score but for some reason weren't included with 10.3.1. Then it will be Actual "Wide Release" with the holiday update because it is a flashy way to end Q4.

So it may be that the holiday update is 2022.44 or something and 2022.40 releases with single stack V11 activated for everyone. With the holiday update FSD beta releases stop (potentially forever).

It's fun to guess.

 

[AlanSubie4Life]

Are you saying something different to that?

Yes.

You may be right that Elon actually means wide release when he says wide release (what you describe in post above is wide release).

However I expect that people will still need to request access. I may be wrong. We’ll see.

If they do it without any request requirement to all buyers (in a particular region), it is wide release, yes.

Some people like @Daniel in SD thought it would never happen! He even had a poll. We’ll see.

 

[Supcom]

Yes.

You may be right that Elon actually means wide release when he says wide release (what you describe in post above is wide release).

However I expect that people will still need to request access. I may be wrong. We’ll see.

If they do it without any request requirement to all buyers (in a particular region), it is wide release, yes.

Some people like @Daniel in SD thought it would never happen! He even had a poll. We’ll see.

I don't think that anyone will need to request city streets FSD once it is released wide at the end of the year. It will still be 'beta' like all of Tesla ADAS features, but it will no longer be 'early access beta'. No safety score will be required. All you will need is have an FSD license (purchase or subscription) and to enable it like one does currently for AP or NOA.

I expect that strikes will be replaced with the AP jail that current production users have.

It will be interesting to see if the existing beta tester program will remain for testing advance improvements, but I expect it wont.

In any event, I'm looking forward to seeing the first you tube videos of beta 11.X to see how it works.

 

[thewishmaster]

While it is within the realm of possibilities that V11 could roll out to all FSD users before the end of the year here is a FAR more likely theoretical scenario timeline for such a LARGE update that will likely need lots of work.

No way would I count on V11 making a full wide releases as a holiday release (unless you count St Patrick's Day) .

• V11.0 release to influencers on Dec 16th
• Several bugs identified
• V11.1 released to influencers on Jan 6th
• V11.1 released to a few Beta users on Jan 13th
• Rollout stopped, likely more bugs
• V11.2 released on Jan 20th.
• Slow rollout over the next 2 weeks to Beta users
• V11.3 released on Feb 17th
• After a couple of weeks in March Tesla decides to allow it to go to all FSD users

You're talking about the same people that released Smart Summon in its existing state

 

[AlanSubie4Life]

I don't think that anyone will need to request city streets FSD once it is released wide at the end of the year. It will still be 'beta' like all of Tesla ADAS features, but it will no longer be 'early access beta'. No safety score will be required. All you will need is have an FSD license (purchase or subscription) and to enable it like one does currently for AP or NOA.

We’ll see!

That would be a very aggressive expansion indeed. (What percentage of FSD buyers have FSD Beta?)

I think they would need to be extremely confident to go that wide that quickly. With a brand new highway stack too! It would boggle my mind, I have to say. I just can’t comprehend it.

Anyway, as I said, we will see.

 

[animorph]

My Model X camera replacement isn't scheduled until 12/23 so a wide release after that would be nice.

 

[AlanSubie4Life]

We’ll see!

That would be a very aggressive expansion indeed. (What percentage of FSD buyers have FSD Beta?)

I think they would need to be extremely confident to go that wide that quickly. With a brand new highway stack too! It would boggle my mind, I have to say. I just can’t comprehend it.

Anyway, as I said, we will see.

I should add that the removal of the report button provides a hint that I am wrong. However, I repeat that I would be completely befuddled by an actual wide rollout to everyone by the end of this year (2022). It just is hard to fathom, given where FSD currently is. This is not like Smart Summon, at all - this is something that people will actually try to use!

Maybe between this and Twitter, Elon is demonstrating that he has no more f***s to give. Ending 2022 with a bang.

 

[Supcom]

I should add that the removal of the report button provides a hint that I am wrong. However, I repeat that I would be completely befuddled by an actual wide rollout to everyone by the end of this year (2022). It just is hard to fathom, given where FSD currently is. This is not like Smart Summon, at all - this is something that people will actually try to use!

Maybe between this and Twitter, Elon is demonstrating that he has no more f***s to give. Ending 2022 with a bang.

Prepare to be befuddled. I really believe that something is going out to all FSD vehicles by the end of the year. Musk has set that as a big goal for Tesla. If 11.X is not ready, then something like 10.69.3.X will roll out. I expect that the only big difference will be single stack or not. the 11.X city streets is almost certainly very similar to 10.69.3.1 and not likely to have a functional breakthrough in the next month beyond replacing NOA (and AP?). So, as I've said elsewhere, there should not be much downside in releasing something without single-stack to everyone.

 

[AlanSubie4Life]

So, as I've said elsewhere, there should not be much downside in releasing something without single-stack to everyone.

I think there is a lot of downside to Tesla of doing this. Again, think about the expansion. True that many might not turn it on, but the viral nature of such a release means that many will.

And given that people had such a hard time getting a safety score of 99 to 100 (not very difficult!), and given the "interesting" behaviors documented elsewhere here in many videos, I do have concerns about the wider rollout to people who are not expecting those behaviors (attempted running of red lights, sudden steering wheel movements mid-turn, etc.).

Perhaps Tesla knows people just don't use it all that much, or have data that shows that there's little risk in rolling out to drivers with less motivation to keep safety score high (they should have got that information from recent safety score expansions). Still, I think it's a different cohort of people who have not even requested FSD Beta. So I don't think Tesla can really predict anything about how they will use it.

And do they really have enough miles on FSD Beta to show anything significant? Can they show it is actually safer than the alternative? Elon has said that is the metric (paraphrasing).

 

[Dewg]

I think there is a lot of downside to Tesla of doing this. Again, think about the expansion. True that many might not turn it on, but the viral nature of such a release means that many will.

And given that people had such a hard time getting a safety score of 99 to 100 (not very difficult!), and given the "interesting" behaviors documented elsewhere here in many videos, I do have concerns about the wider rollout to people who are not expecting those behaviors (running red lights, sudden steering wheel movements mid-turn, etc.).

Perhaps Tesla knows people just don't use it all that much, or have data that shows that there's little risk in rolling out to drivers with less motivation to keep safety score high (they should have got that information from recent safety score expansions). Still, I think it's a different cohort of people who have not even requested FSD Beta. So I don't think Tesla can really predict anything about how they will use it.

And do they really have enough miles on FSD Beta to show anything significant? Can they show it is actually safer than the alternative? Elon has said that is the metric (paraphrasing).

My guess is that some people will turn it on and use it, and if their cars are not properly maintained (clean, calibrated, wheels aligned, GPS accurate, etc.), Autosteer on City Streets will be a disaster for them. And very likely those people will turn it off and never turn it on again, cursing Tesla's name privately or on forums. The end result will be the same - well maintained cars will enjoy the new features, and the rest won't use it after it fails the first time.

 

[AlanSubie4Life]

My guess is that some people will turn it on and use it, and if their cars are not properly maintained (clean, calibrated, wheels aligned, GPS accurate, etc.), Autosteer on City Streets will be a disaster for them. And very likely those people will turn it off and never turn it on again, cursing Tesla's name privately or on forums. The end result will be the same - well maintained cars will enjoy the new features, and the rest won't use it after it fails the first time.

I think it is WAY too good for that outcome. This is no Actually Dumb Smart Summon.

It's actually quite good at driving on a straight road with traffic lights. Not particularly smooth or silky, but certainly it does acceptably well for someone who wants to be doing something else. That is, until it runs a red light for no apparent reason (I haven't seen anyone run a run light on video, but you can see my video where it could have done exactly that (we'll never know)). Or whatever malfunction you want to pick; there are many!

And while the driver monitoring is actually quite good, it's not all that hard to fool the system and do whatever else you want to be doing. Certainly it's possible to drive in fairly complex scenarios without paying any significant attention to what is going on. And with a wide rollout, I don't see them permanently locking people out (though perhaps they will maintain that for now - that would be one way to mitigate risk - they could even lower the number of strikes to three for everyone, which would be great).

For someone who is drunk it would probably be particularly good, and no need for attention monitoring defeats. It is unclear to me whether or not it is better than a drunk driver (of course it would depend on the BAC). Regardless, not a great look to crash when drunk on FSD, even if that accident "was going to happen anyway" - that argument is not going to fly.

cars are not properly maintained (clean, calibrated, wheels aligned, GPS accurate, etc.), Autosteer on City Streets will be a disaster for them

Not sure what any of these things have to do with Autosteer on City Streets function. Nearly all cars meet the requirements to have Autosteer work correctly. If not, we'd see tons of reports about people having problems with these things (instead, we just see a few reports of these listed issues).

 

[Supcom]

My guess is that some people will turn it on and use it, and if their cars are not properly maintained (clean, calibrated, wheels aligned, GPS accurate, etc.), Autosteer on City Streets will be a disaster for them. And very likely those people will turn it off and never turn it on again, cursing Tesla's name privately or on forums. The end result will be the same - well maintained cars will enjoy the new features, and the rest won't use it after it fails the first time.

For what it's worth, my car has not been washed in over a month, it's been at least two months since I did a camera calibration and my one year+ old car has never had the wheels aligned. I use FSD beta on virtually every drive and do not see performance worse than what others report here. In many ways it's better, but I realize that complaints are more often expressed than compliments! So, I expect few FSD experience disasters due to that.

I do agree that there will be a lot of discouraged users as Tesla does a poor job of managing customer expectations. And that's putting it mildly!

 

[Dewg]

We'll see. I'll keep my fingers crossed.

 

[thewishmaster]

So........is next month in Elon speak supposed to be more optimistic or more pessimistic than the last v11 "timeline confirmation" of before T-Day?

View attachment 876613

On the brink of Tuesday, with 10.69.3.1 ostensibly not in wide release yet (stalled around ~480 cars on teslafi), how many more cars do we reckon will get v11?

 

[RowdyMY]

Prepare to be befuddled. I really believe that something is going out to all FSD vehicles by the end of the year. Musk has set that as a big goal for Tesla. If 11.X is not ready, then something like 10.69.3.X will roll out. I expect that the only big difference will be single stack or not. the 11.X city streets is almost certainly very similar to 10.69.3.1 and not likely to have a functional breakthrough in the next month beyond replacing NOA (and AP?). So, as I've said elsewhere, there should not be much downside in releasing something without single-stack to everyone.

I really hope this isn't the case.

I-95 was FUBAR yesterday and I took 301 across the Potomac. Had to switch to my non FSD profile 3x just to get the dang thing to stay out of the left lane or, ironically, from trying to dive into the decel lane on a right shoulder.

FSDb as a one size fits all is absolutely not ready for a prime time release. I love this car and I want it to be, but it's not.

Eta - for those not aware, 301 is the only alternative to I95 in NOVA. It is also not a limited access/NoA highway and uses FSDb despite it's mostly 60 MPH speed limit.

 

[powertoold]

New video if you want to geek out over FSD development. Basically, Tesla currently uses over 5000 FSD computers to do their unit tests for new builds.

Each FSD computer is fed either real or simulated clips, and then for the simulated clips, the FSD chip consumes the video stream, outputs the controls, and then the computer farm generates new simulated images based on the controls outputted by the FSD chip. I'm not sure how real clips are used.

Overall, kind of a boring video.

 

[gsmith123]

It will be interesting to see if the existing beta tester program will remain for testing advance improvements, but I expect it wont.

I hope they do, and I could see a lot of reasons to do so. They have a robust incremental rollout strategy for FSD updates (developers, all Tesla employees, small external beta group / long time beta testers, newer beta testers, all users). This is exactly how you want to roll out software that is complex to test and is safety critical.

It also gives them user groups to do AB testing on. I'm not sure Tesla has ever confirmed this for sure, but there's a lot of anecdotal experiences where different testers on the same version see a surprisingly large variation in performance (consistent with some runtime tuning that is different).

After the full public wide release of FSD for city streets and single stack highway AP, they've still got a lot of development to go. Offhand:

- Parking lots
- Driveways and garages (really anything that's not a public road)
- Any kind of maneuvers that require backing up
- Inclement weather
- Slippery surfaces (snow and ice)
- Locations beyond US + Canada
- and just incremental improvements on everything FSD can do today to make it more reliable and work towards L3/L4 operation
- (maybe) some kind of HW4 upgrade with cross traffic cameras further forward on the vehicle (presumably would require some retraining and testing)