FSD v12.x (end to end AI)

[rlsd]

My car is fine. It has modern extremely capable hardware, capable of full autonomy: HW3.

Drive on a road at 50-60mph with traffic lights every half mile or so and report back.

I am curious if HW4 does better, actually, since I believe part of this could potentially be usable perception range. But I do think it is more likely a software limit, not perceptoon.

Everything is possible. I would not rule out camera defect or other component defect. If you continue to have problem I think you should complain with Tesla service to check your car.

 

[FSDtester#1]

Everything is possible. I would not rule out camera defect. If you continue to have problem I think you should complain with Tesla service to check your car.

If I had a nickel for every time someone suggested he got his car checked....

 

[LowlyOilBurner]

Can Tesla enable video in motion, like YouTube or Netflix while waiting for some turns to be completed on V12.3? Some of these turns are taking waaaaaaay to long to complete. 11.4.9 was better in some of these instances. My wife even commented “I think we’re going to be later for dinner if this is the way it drives now”. Assuming the slow driving speeds and the ridiculously slow times it takes to complete turns when there is ZERO cars at some intersections…

 

[Todd Burch]

Why are we still getting the disengagement notice at the end of the navigation?
They turned this off when we were close to our destination, now it's back on..
I just say, "I am home, no problems Tesla team" when I get the "What happened"
Wasted data for them to sift through.

I assume that if you don’t record a message, it won’t send the report.

 

[hobertg]

Looks like a new wave started this evening

 

[Supcom]

Can Tesla enable video in motion, like YouTube or Netflix while waiting for some turns to be completed on V12.3? Some of these turns are taking waaaaaaay to long to complete. 11.4.9 was better in some of these instances. My wife even commented “I think we’re going to be later for dinner if this is the way it drives now”. Assuming the slow driving speeds and the ridiculously slow times it takes to complete turns when there is ZERO cars at some intersections…

Maybe they'll enable "Walk on ahead" along with ASS and Banish. Then, you can let the car continue on to meet you at the destination when it's time to return home.

 

[Supcom]

Looks like a new wave started this evening

Dang! I got in the second wave of 12.3. I was hoping to continue my run of good luck. Oh well. I got to drive 12.3.1 on two service loaner cars this week, so that should tide me over until my turn comes.

 

[JHCCAZ]

Which by the way, I had to find out if it is legal to floor it at a light.
Every officer I asked said go for it.

I'm sure it depends on the situation and the individual officer. But street racing is technically illegal in most places even if you don't exceed the limit. And even if you're alone, there's something called Display for Acceleration that you can be ticketed for.

But obviously this was a super common thing back in high school, and I can't say I've lived a completely clean life at traffic lights. It's funny, now that I have probably the quickest car I've ever owned (even if not Performance or Plaid) I feel no need to prove it.

 

[JHCCAZ]

I assume that if you don’t record a message, it won’t send the report.

I believe that the disengagement is flagged, and especially right now is probably uploaded. Look at how much your car is uploading since you got 12.3, whether or not you've been adding audio commentary after the disengagements.

I agree with those who say that Tesla would really help the community, and presumably themselves, if they would explain exactly what they are looking for in the What Happened reports.

 

[AlanSubie4Life]

there are fewer accidents per million miles on FSD Beta than for Teslas without active safety features.

Driving on FSD Beta, on the roads that FSD Beta is typically driven on, with the disengagements that drivers have been making, is safer than not [using FSD Beta on these same roads and same conditions that FSD Beta is typically driven on]

(Words in brackets added by me.)
If people can't see that these are very obviously different statements, and not equivalent at all, I am not sure I can help you. The first statement is true, per Tesla's data. The second statement clearly does NOT follow from the second. It may be true. We don't know.

Why? You won’t believe it.

I believe all the data published by Tesla to date. Why would I not believe additional data?

I drive on a lot of roads at 55-60mph with traffic lights every few miles. HW4 does not do better

Your descriptions sound like v11 though.

I have [HW4] and the braking doesn't bother me at all.

I have a HW3 car and drive on just such a road daily. 11.4.9 was extremely good. 12.3 has been virtually flawless.

All very interesting.

On my way home today on Mira Mesa I had zero interventions for rough stopping. The traffic was heavier and the flow was such that there was minimal opportunities for brake slamarama. There was perhaps a bit more light. It's really just luck of the draw and more traffic generally helps unless you're behind someone who does not have good flow.

If you continue to have problem I think you should complain with Tesla service to check your car.

I will do no such thing. That black hole is not to be approached unless absolutely necessary.

It very well may - that’s a natural human tendency so it may have been trained into the system.

I really like the way v12 moves around in lanes based on positions of other traffic. It actually maintains buffers sometimes, similar to a human. It's not perfect, but it does drift around very nicely. Sometimes it is exaggerated, and there are general issues with the vehicle taking the correct line on some roads. But nothing egregious so far (I'm sure it happens).

 

[AlanSubie4Life]

0 for 1 on unprotected lefts on 12.3.1, @Daniel in SD. The hangs by a mere thread.

We'll see if he goes 9/9 on the next 9! I'm sure it was just the rain; those cars were super hard to see with their headlights and all.

No normal human could perform this turn in these conditions. This is a highly skilled human trained to land on aircraft carriers in a nuclear-capable (depth charges - seems like overkill?) plane resembling a flying Hoover vacuum (S-3 Viking).

Also failed here (doesn't count):

 

[TK211X]

Looks like a new wave started this evening

Wish I could see the engineers face when the hand-full of updates he sent to vehicles owned by Mia Khalifa worshipping gold diggers who got a Tesla from their sugar daddy ignore the update notification in the sea of Tik Tok notifications.

 

[kabin]

0 for 1 on unprotected lefts on 12.3.1, @Daniel in SD. The hangs by a mere thread.

We'll see if he goes 9/9 on the next 9! I'm sure it was just the rain; those cars were super hard to see with their headlights and all.

No normal human could perform this turn in these conditions. This is a highly skilled human trained to land on aircraft carriers in a plane resembling a flying Hoover vacuum.

Also failed here (doesn't count):

A bit later (@20:45) V12.3.2.1 willingly tailgated a lead vehicle without urgency of applying sufficient regen/brakes. I don't recall V11 ever doing that

 

[TK211X]

I wanted to put more than on reaction to this. I went with the crowd and did the laughing guy. You're really winning lately.

Well, hey, if you ever need happiness go to the rabbit. I mean catch the rabbit.

…..but if you get caught naked at a rabbit farm, do not mention my name. At that point they’re going to start noticing a pattern.

 

[Daniel in SD]

0 for 1 on unprotected lefts on 12.3.1, @Daniel in SD. The hangs by a mere thread.

I feel judging the performance in the rain is unfair. I don’t think the test drivers have ever been there in the rain. Though I still feel that I’ve got a good chance on the next ten turns. It rarely rains in Florida.

 

[AlanSubie4Life]

I feel judging the performance in the rain is unfair. I don’t think the test drivers have ever been there in the rain. Though I still feel that I’ve got a good chance on the next ten turns. It rarely rains in Florida.

Ok! 0/0 it is! Feeling confident. v12 is superbly good, but not 9/10 good. It has been quite good indeed on my turn. Probably at least 50% success rate.

 

[AndrewZ]

He’s an optimist!
If you assume that necessary disengagements drop by 50% every couple weeks then unsupervised FSD will be ready this year.

He’s a man of extremes for sure, his tweet history is full of other things that we can maybe just say he’s not so optimistic about lol

 

[Todd Burch]

I have a 2022 and it can take almost 5 mins to wake sometimes, then other times it only takes 30 seconds

Even my 2012 never did that.

Like an anal cyst, that is not normal and you should get it looked at.

 

[Todd Burch]

I disagree. A fatal crash requires a higher level of driving failure than just a crash. As in, the type to result from inattention or a series of poor choices.
It requires completely missing either a stationary or moving object with a high differential speed.

Not speeding alone cuts risk down. As does not drinking. FSD is good at both of those.
https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/813473

A good example:

A lot of fatalities happen on two-lane roads where one driver looks down at his/her phone or otherwise gets distracted, swerves into an oncoming lane, and has a head-on with the oncoming lane.

I’ve only caught FSD texting its girlfriend once in the last 4 years.

 

[Todd Burch]

That's meaningless without data showing FSD is safer. And we don't have it, so cannot judge it.

I know.

Yeah. To be clear, as long as my car works at 3rd-party NACS DC chargers with the Tesla connector without needing a charge port update, I am not complaining, and I would agree Tesla won, by adopting CCS protocol. Otherwise, I would have to say my car has been left in the past because it uses the obseleted Tesla-proprietary standard. It’s not clear to me what will happen (it would make no sense to not support older Teslas, since Tesla does). I would hate to plan a trip using a remote NACS charger as the only option and not be able to use it!



No, there is zero data. Nothing has been published. It’s very frustrating! (Tesla does continue on occasion to publish statistics with obvious massive selection bias.)

Tesla updates and publishes safety statistics every quarter:

https://www.tesla.com/VehicleSafetyReport

You are free to go all tinfoil hat and say they are fudging the numbers or you can collect them yourself.