Poll: When will FSD V12 be in wide release?

[Daniel in SD]

That’s not what I said at all.

It was a joke. What I’m saying is that sorting through all the disengagement data to figure out which ones were real safety issues is not easy. You have to predict what would have happened had the safety driver not disengaged. In an L3-5 system with no safety driver it is very easy to find real safety issues (because there is a collision while the system is engaged).

 

[rlsd]

I think there are many factors for Tesla to indicate that FSD has L3 or not:

1. FSD can actually perform L3.
2. How many vendors have L3 for self driving.
3. How many accidents involved with L3 cars, how many lawsuits and settlement costs.
4. How many cars sold by Tesla with FSD.

 

[gsmith123]

What I’m saying is that sorting through all the disengagement data to figure out which ones were real safety issues is not easy. You have to predict what would have happened had the safety driver not disengaged. In an L3-5 system with no safety driver it is very easy to find real safety issues (because there is a collision while the system is engaged).

You also don’t get any disengagements for all the problems that don’t result in crash. IMO, that’s most of the problems with FSD at this point.

 

[Daniel in SD]

You also don’t get any disengagements for all the problems that don’t result in crash. IMO, that’s most of the problems with FSD at this point.

Lets say Tesla has an FSD build they think could release as an L3 system.

I was talking about a hypothetical system that would be safe enough to release as an L3 system. At that point you've got to simulate counterfactuals for thousands of disengagements before you find one that was necessary (that's what Waymo had to do).
There are so many complaints in this forum about FSD requiring you to pay attention to the road. I wouldn't underestimate how valuable non-beta FSD would be.

 

[AlanSubie4Life]

Currently FSD is so far from L3 in any ODD that I'm not sure why we're discussing it.

Just because you've never seen it do anything wrong in slow stop & go traffic doesn't mean it's anywhere near good enough for Tesla to take on liability. The bar is super high. ~100k participants really increases likelihood of seeing an issue that at an individual level you have never seen. You could drive for years and not see an issue that shows up on the first day Tesla rolls out the feature.

And what's the upside for Tesla? Are they going to charge FSD owners additional money for the privilege of L3? That would go over like a lead balloon.

I would think Tesla has a pretty good idea of where they're at.

 

[xG35]

This thread is still alive? I have bought fsd 3 times on 3 teslas… let’s face it, it ain’t coming anytime soon.

It’s great for what it does, but that full self driving promise, is pretty distant unless it finds its GPT moment.

 

[boonedocks]

This thread is still alive? I have bought fsd 3 times on 3 teslas… let’s face it, it ain’t coming anytime soon.

It’s great for what it does, but that full self driving promise, is pretty distant unless it finds its GPT moment.

Don't feel alone. I have bought it 3 times as well but did not on my 2021 Refresh S. I did however pay$200 a month for almost a year - call me an IDIOT . Basic AutoPilot is definitely the way to go until there is ACTUAL FSD. Maybe two more S' for me and we'll see

 

[JulienW]

This thread is still alive? I have bought fsd 3 times on 3 teslas… let’s face it, it ain’t coming anytime soon.....

The thread is about V12 coming and not L3 or L4 Full Self Driving. While it is still probably a month or so away it does appear to be coming quicker than most (including me) believed.

 

[eevee-fan]

This thread is still alive? I have bought fsd 3 times on 3 teslas… let’s face it, it ain’t coming anytime soon.

It’s great for what it does, but that full self driving promise, is pretty distant unless it finds its GPT moment.

3 times? You know what they say in TN...

 

[xG35]

3 times? You know what they say in TN...

I agree.. freaking suckers.. and I am a tesla investor… but this is just dreadful at this point.

 

[StealthLSU]

Currently FSD is so far from L3 in any ODD that I'm not sure why we're discussing it.

Just because you've never seen it do anything wrong in slow stop & go traffic doesn't mean it's anywhere near good enough for Tesla to take on liability. The bar is super high. ~100k participants really increases likelihood of seeing an issue that at an individual level you have never seen. You could drive for years and not see an issue that shows up on the first day Tesla rolls out the feature.

I disagree. I think a big problem is by having an L3 situation, you would have to limit the other parts of FSD.

For instance, my understanding of the BMW L3 is it goes under 37 mph and follows a car but won't change lanes.

So for Tesla to copy that, while L3 is activated, all other FSD systems(auto lane change or lane change by blinker) would have to be disabled since those wouldn't fall under what they would approve for L3.

Because of all that, even from my perspective it seems like it is more trouble than its worth. But who knows, maybe they do some version of it for the headlines.

 

[eevee-fan]

For instance, my understanding of the BMW L3 is it goes under 37 mph and follows a car but won't change lanes.

So... a Lemming mode? Will it follow a car off the cliff? LOL

 

[AlanSubie4Life]

it seems like it is more trouble than its worth.

those wouldn't fall under what they would approve for L3.

I mean…I guess this is my point. The ODD would have to be vanishingly small for Tesla to approve it and take liability.

There is probably some ODD they would be comfortable with. Maybe. But it would not be useful.

I also don’t think they would be comfortable with 37mph with lead traffic and no changing lanes! That is VERY tough to get right.

 

[PianoAl]

What I’m saying is that sorting through all the disengagement data to figure out which ones were real safety issues is not easy.

It would help Tesla if they suggested wording for disengagement reports. For example, if the car does something really bad, I start me report with, "Major bug: ..."

Suggested phrases, minor problem, inconvenience, dangerous situation, almost crashed, annoyed other drivers, lane drift, phantom breaking, failure to yield, wrong speed, etc., might help.

 

[JulienW]

Now that we know all that bet on January LOST BIG TIME do we have any February true believers?

 

[AlanSubie4Life]

Now that we know all that bet on January LOST BIG TIME do we have any February true believers?

January wasn't a choice.

I expect before July though I think there's a real chance I could be wrong.

 

[Supcom]

Now that we know all that bet on January LOST BIG TIME do we have any February true believers?

I still say April, at the earliest.

 

[JulienW]

January wasn't a choice.

I expect before July though I think there's a real chance I could be wrong.

There are a few post of some were saying imminent or January release. That is what I'm referring to.

 

[Supcom]

There are a few post of some were saying imminent or January release. That is what I'm referring to.

Hope springs eternal - even when it's hopeless.

 

[eevee-fan]

Elon's Brother, Kimbal Musk, Comments on When Full Self-Driving Will Be Achieved

Discover Kimbal Musk's take on Tesla's journey to autonomy, and the future of driving in this insightful overview.

www.notateslaapp.com

"However, Musk is candid about the challenges ahead, particularly the regulatory hurdles that Tesla must overcome. He predicts that full autonomy could be achievable within five years, contingent upon navigating the intricate landscape of government regulations and ensuring the technology meets rigorous safety standards."

So... 10 years or longer...