Elon: FSD Beta tweets

[kabin]

Yeah, maybe. In the end the cameras will obviously need more than that range, and they probably do already. At 10:00 in the video above you can see it has visibility out to 100m. But then the vehicle is immediately tripped up by occlusion - they really need to reliably identify all the important objects in the scene, immediately, and remember them, just like a human! Another capability unlock which will be enabled by mastering this turn.

I recall Chuck saying he doesn't have access to the B pillar camera data so he simulates the B camera view with another camera placed on the roof and in alignment with the B pillar camera. It's anyone's guess what that camera's specs are.

 

[GlmnAlyAirCar]

I’ve driven about 15k miles in total in the last year, maybe 90% on AP/FSD.

The 3.8 mile stretch I referenced in an earlier post is something I’ve done basically every workday since September of 2021 when I got into the program. FSD has never been able to complete that straight shot down a major thoroughfare in Los Angeles that probably sees 5k Teslas a day. To this day it cannot make it straight through two of the intersections on green lights on its own. At one point it gets all the way over into a turn-only lane onto the 110 freeway instead of proceeding through the green light ahead of it.

The car cannot go from my office parking structure down a well-marked road for 500 feet and make a left with a stop sign onto another well-marked street and proceed another 500 feet. That little situation, without traffic, involves three full disengagements. I’ve been doing that for 11 months too.

Last week my car uploaded nearly 50GB of video. FSD is just…it just doesn’t work. It’s a novelty. View attachment 837255

Interesting. The one factor that I found does lead to more disengagements is traffic density. There seems to be a point from which it will go from great performance to terrible if density gets beyond a certain point. By density I’m not only talking about the amount of nearby traffic but also it’s speed. Gridlock generally does much worse than steadily moving, heavy traffic. But my point stands that on my regular routes, even in very dense traffic, it does well.

even though there are many teslas operating over your route, how many are actually running FSDb consistently over the entire stretch every day? Even though you may be, I suspect the fast majority are not.

 

[jabloomf1230]

I recall Chuck saying he doesn't have access to the B pillar camera data so he simulates the B camera view with another camera placed on the roof and in alignment with the B pillar camera. It's anyone's guess what that camera's specs are.

Here's my favorite Chuck Cook video, where he first mounted GoPros on the left B pillar and the left headlight in order to roughly compare the two views.

 

[Smh2210]

how many of you are still believing ANYTHING Elon says about FSD these days. ANYTHING

I believe in the King of FSD.

 

[Smh2210]

Interesting. The one factor that I found does lead to more disengagements is traffic density. There seems to be a point from which it will go from great performance to terrible if density gets beyond a certain point. By density I’m not only talking about the amount of nearby traffic but also it’s speed. Gridlock generally does much worse than steadily moving, heavy traffic. But my point stands that on my regular routes, even in very dense traffic, it does well.

even though there are many teslas operating over your route, how many are actually running FSDb consistently over the entire stretch every day? Even though you may be, I suspect the fast majority are not.

I think we users disengage more frequently when things get tricky and we intercede to be safe. I also assume (I know.....), Tesla staffers let the car go and see how it does with a longer leash than 'us', thus higher success.

 

[myshenaini]

I think we users disengage more frequently when things get tricky and we intercede to be safe. I also assume (I know.....), Tesla staffers let the car go and see how it does with a longer leash than 'us', thus higher success.

Perhaps staffers are covered by insurance from the company so even if they got into an accident, it won't cost them.

 

[Daniel in SD]

Perhaps staffers are covered by insurance from the company so even if they got into an accident, it won't cost them.

That doesn't protect them from criminal charges though. Intentionally letting the car do something dangerous would not look good in court.
Also I don't understand the point of "seeing what will happen." Either the car is driving safely or it isn't. The whole point of testing is to disengage when it's doing something wrong.

 

[bobsmy]

I think we users disengage more frequently when things get tricky and we intercede to be safe. I also assume (I know.....), Tesla staffers let the car go and see how it does with a longer leash than 'us', thus higher success.

I think we all have different length leashes based on how we normally drive. One interesting consequence of interceding later is it is more likely to generate a warning ( dreaded red wheel). All but one of my warnings have been the result of me taking over when the car was already in an unsafe situaation.

 

[Jeff N]

I’ve noticed that some problematic areas of my regular routes are handled just fine if I lower the speed. At 30 mph it might have trouble with an awkward lane splitting off to the side and converting into a turn-only lane but at 15 mph it handles it just fine.

I’ve let it mess up at regular speed and reported it multiple times so I feel fine about manually trimming down the speed via the steering wheel controls to help it out now.

When 10.13 finally arrives I will go back to letting it do it’s thing at full speed.

 

[AlanSubie4Life]

OMG, they will solve the left turn in just two weeks.

Unfortunately 10.69 is not a revision we are familiar with.

https://twitter.com/x/status/1556181455570714624

 

[JulienW]

OMG, they will solve the left turn in just two weeks.

Unfortunately 10.69 is not a revision we are familiar with.

https://twitter.com/x/status/1556181455570714624

....or remove the reference from the new release notes.

Also had we been getting FSD Beta updates at the original pace Elon stated, every 2 weeks then we would already be on 10.69 by now.

 

[Smh2210]

I’ve noticed that some problematic areas of my regular routes are handled just fine if I lower the speed. At 30 mph it might have trouble with an awkward lane splitting off to the side and converting into a turn-only lane but at 15 mph it handles it just fine.

I’ve let it mess up at regular speed and reported it multiple times so I feel fine about manually trimming down the speed via the steering wheel controls to help it out now.

When 10.13 finally arrives I will go back to letting it do it’s thing at full speed.

Agree on speed, it has taken me over the double yellow a few times, and I have reported it.

 

[GranularHail]

OMG, they will solve the left turn in just two weeks.

Unfortunately 10.69 is not a revision we are familiar with.

Elon said in the shareholder meeting presentation last week something to effect of 10.13 is getting so good it should really be called 10.69. See about 8 minutes into the presentation:

So I think he's saying what we knew as 10.13 would be released in basically "two weeks"

 

[diplomat33]

OMG, they will solve the left turn in just two weeks.

Unfortunately 10.69 is not a revision we are familiar with.

https://twitter.com/x/status/1556181455570714624

10.69 is the same update as 10.13. Elon just renamed 10.13 to 10.69 because he likes silly references to 69 and 420.

 

[Daniel in SD]

Nice, but does this mean it’s going to go down and make a u-turn instead of the left?

 

[aronth5]

10.69 is the same update as 10.13. Elon just renamed 10.13 to 10.69 because he likes silly references to 69 and 420.

Plus Chuck Cook was born in October 1969! The stars are aligning for a good update

https://twitter.com/x/status/1556236824892379138

 

[AlanSubie4Life]

Nice, but does this mean it’s going to go down and make a u-turn instead of the left?

Nope, 100% success on the left turn, as Elon said.

 

[diplomat33]

Plus Chuck Cook was born in October 1969! The stars are aligning for a good update

https://twitter.com/x/status/1556236824892379138

Oct is the tenth month. So maybe 10.69 is simply an easter egg reference to Chuck since the update will supposedly "solve" Chuck's ULT.

 

[Twiglett]

Nope, 100% success on the left turn, as Elon said.

I think he said in the shareholder meeting that they reckon 90% success with "Chucks" left turn.
Unclear on what happens for the other 10% though

 

[AlanSubie4Life]

I think he said in the shareholder meeting that they reckon 90% success with "Chucks" left turn.
Unclear on what happens for the other 10% though

No he said that was where they were at, so they are close to 100% or something like that. I quoted it elsewhere. It was pretty clear from the quote that the goal is 100%. Some issues with that commitment, also covered earlier.

“I think we’re roughly at 90% with your turn…so…we’re almost at 100…so…it’s looking good…so…I’m hopeful it might be next week.”

If that’s not a firm commitment to 100%, I don’t know what is.