FIRMWARE UPDATE! AP2 Local road driving...and holy crap

[u00mem9]

Early adopters should get T shirts!
"I survived AP2"
With a model S or X below it

I hope that stays funny....(cringe)

 

[sdorn]

The general public doesn't even know that Tesla has any form of auto pilot, much less that the hardware was recently changed. Many don't even know that Teslas are electric. I've had plenty of people tell me they like Teslas and are then surprised when I tell them they are electric. Hardly anyone knows anything about auto pilot and are usually blown away when I tell them it can steer itself in certain situations.

 

[Dan3690]

I would buy one.

 

[croman]

Given the limitations of AP2 currently, I haven't been able to use it for more than a 2 or 3 mile stretch in stop and go traffic. From what I've seen so far, I would not say that it reduces stress or fatigue. I do think that once high speed AP2 is enabled that might change, but that is pure speculation. About the worst thing AP2 has done so far for me is not being able to take a rather sharp turn on a highway exit (not really designed for that) and trying to drive the car onto the shoulder of the interstate when the sun blinded the camera and it lost the lane marker (which only happened once right after AP2 was first enabled).

I've used highway AP2 for 10 miles repeatedly. Other than false FCWs (which are random as far as I can tell here in Chicago), it is as good as AP1 (except speed). Local autosteer I've managed 5.6 miles on a road that has barely any turns and is 4 lanes (2 in each direction + only lanes that appear and disappear as the center median appears and disappears for intersections). There are no stop signs, only stop lights and I have to disengage if I don't have a lead car but I do this commute everyday and can attest that local autosteer is not as bad as people say but its clearly also not to be trusted.

I 100% agree with @Papafox that if you lightly rest your hands on the wheel it will not nag you and you can feel what the car is doing. I keep a close watch on local autosteer but highway autosteer has been really great for me.

Even local autosteer makes my commute way less stressful because as long as there are no curves, it doesn't try to kill me.

 

[Deans]

Early adopters should get T shirts!
"I survived AP2"
With a model S or X below it

I hope that stays funny....(cringe)

I just hope that stays true.

 

[disagree]

I've used highway AP2 for 10 miles repeatedly. Other than false FCWs (which are random as far as I can tell here in Chicago), it is as good as AP1 (except speed). Local autosteer I've managed 5.6 miles on a road that has barely any turns and is 4 lanes (2 in each direction + only lanes that appear and disappear as the center median appears and disappears for intersections). There are no stop signs, only stop lights and I have to disengage if I don't have a lead car but I do this commute everyday and can attest that local autosteer is not as bad as people say but its clearly also not to be trusted.

I 100% agree with @Papafox that if you lightly rest your hands on the wheel it will not nag you and you can feel what the car is doing. I keep a close watch on local autosteer but highway autosteer has been really great for me.

Even local autosteer makes my commute way less stressful because as long as there are no curves, it doesn't try to kill me.

No offense intended, but what you describe is pretty much "if your local road is like stop and go highway traffic under 35mph, then EAP is not as bad as in other circumstances."

On the upside, local roads have fewer overheads signs!

 

[NerdUno]

Early adopters should get T shirts!
"I survived AP2"
With a model S or X below it

I'm reminded of the old saying about pioneers and think I've found an appropriate image to remind everyone of their AP2 experience:

 

[croman]

No offense intended, but what you describe is pretty much "if your local road is like stop and go highway traffic under 35mph, then EAP is not as bad as in other circumstances."

On the upside, local roads have fewer overheads signs!

None taken and your pithy summation is true but it appears Tesla is figuring it out. I think it's clearly not what Tesla led buyers to believe but they aren't completely failing either. I am skeptical that they will deliver on their most recent statements but it's possible unlike their December expectations.

 

[182RG]

Early adopters should get T shirts!
"I survived AP2"
With a model S or X below it

 

[_jal_]

The general public doesn't even know that Tesla has any form of auto pilot, much less that the hardware was recently changed. Many don't even know that Teslas are electric. I've had plenty of people tell me they like Teslas and are then surprised when I tell them they are electric. Hardly anyone knows anything about auto pilot and are usually blown away when I tell them it can steer itself in certain situations.

I had a guy argue with me that they have a gas motor as a backup. People. They're the worst

 

[Helmuth]

I had a guy argue with me that they have a gas motor as a backup. People. They're the worst

I had one guy asking me if it's a 6 or 8 cylinder because it runs so smooth...

 

[Zetopan]

I had one guy asking me if it's a 6 or 8 cylinder because it runs so smooth...

Depending on which model S you have, there is are definitive answers for such naïve questions. For a P90D(L) the answer is 7104 and for a P100DL the answer is 8256, etc. (see link to cell count). Also be sure to tell them that the very high cylinder count is why it runs so smoothly.
LINK: https://teslatap.com/undocumented/

 

[howardc64]

Am in the tech user interface field and been looking at AI based UI like talking machines (Siri+Echo+copies) and self driving cars, it seems these AI's trained pattern recognition has the ability to provide a good demo under good conditions. But when conditions are more ambiguous, the AI has to give up. For humans users to trust an UI, it has to work under nearly all conditions.

Test drove a service loaner with AP2 last couple of days. It gets lost on lane markers unless they are good/perfectly visible (not the case in real world on local roads). And attempts to stay in center of lanes on gradual turns which feels like turning late to human drivers as we tend to cut into corners on turns.

Detecting/determining the driving lanes under all conditions will be quite a challenge if self driving car developers ever hope to get human out of the car.

 

[NerdUno]

If you haven't seen this article and you're not on blood pressure medicine, have a look:

CA DMV Report Sheds New Light On Misleading Tesla Autonomous Drive Video - DailyKanban

 

[3Victoria]

If you haven't seen this article and you're not on blood pressure medicine, have a look:

CA DMV Report Sheds New Light On Misleading Tesla Autonomous Drive Video - DailyKanban

From another thread, I answered:

Ugh -- I took a look at the linked 'article', under duress. I would say that Tesla did a remarkable job. The Planner/Follower errors seem like problems with communications between the course 'planning' software, which I assume interfaces with the mapping app, and the driving 'follower' software -- likely because they set it up for this demo. After they had debugged it in October, it looks like they had a clean run through the demo course without any disengagements. It is also pretty clear that they are testing this stuff elsewhere than California public roads, but set up the demo on public roads on purpose, and thus had to do a disengagements report. I am not seeing anything here distressing, other than more FUD from DK, which is a well-known anti-Tesla site and written by a well-known troll.

 

[NerdUno]

@3Victoria People that don't agree with your view of the universe are not necessarily trolls. I'm still searching for a factual basis or citation supporting the assertions you have made in your referenced post. Do you work for Tesla? Otherwise, how would you know "they are testing this stuff elsewhere than California public roads" or that Tesla "had a clean run through the demo course without any disengagements?" You obviously haven't driven a production AP2 vehicle or you would appreciate how specious your claims about AP2's functionality really are.

 

[stopcrazypp]

If you haven't seen this article and you're not on blood pressure medicine, have a look:

CA DMV Report Sheds New Light On Misleading Tesla Autonomous Drive Video - DailyKanban

I discussed this briefly, but this is old news. The report explicitly says it does not include any testing done off public California roads. And to even do that demo there obviously must have been development Tesla done in-house (which would not be reflected on the report). To get an uncut take, I would imagine they did many takes. That is not unusual for advertising (where multiple takes are sometimes spliced together too).

FSD may require a hardware upgrade...

 

[NerdUno]

@stopcrazypp There's a gaping hole in your logic. Because the state of California says that the report does not include any testing performed off public California roads does not mean that Tesla actually did testing off public California roads. If, in fact, they did other testing, then one has to wonder why it took 500 miles on public roads with 177 disengagements to make this brief video. This video was represented to buyers and the public as Tesla Vision, not "advertising." I have a car with Tesla Vision that does less than 1% of what is shown in that video.

 

[stopcrazypp]

@3Victoria People that don't agree with your view of the universe are not necessarily trolls. I'm still searching for a factual basis or reference to the assertions you have made in your referenced post. Do you work for Tesla? Otherwise, how would you know "they are testing this stuff elsewhere than California public roads" or that Tesla "had a clean run through the demo course without any disengagements?" You obviously haven't driven a production AP2 vehicle or you would appreciate how specious your claims about AP2's functionality really are.

Obviously you do not know the history between the author (Niedermayer) and Tesla. He was famous/infamous for entertaining the long term suspension "Keef" troll and which Tesla called him out for. There is obviously bad blood between him and Tesla.
A Grain of Salt

As for the demo course, you can just look at the video. As for AP2 (presumably EAP from what you are implying), this is not about AP2 (EAP), but rather about the FSD software, which so far has not been released in public hands.

 

[NerdUno]

@stopcrazypp Perhaps we're discussing different videos. Here's the six-minute version at actual speed. Leaving the mapping functionality aside, well over half of the video showcases AP2 functionality that is not present on current Tesla production vehicles sold with AP2.

Are you suggesting that 500 miles with 177 disengagements on public streets would be normal behavior to film a six-minute video even with the off-site preparation that you assume took place? That's even more scary.