Tesla replacing ultrasonic sensors with Tesla Vision

[Boza]

Not worth my time. Stopped buying Teslas in 2018 so it doesn't affect either of my current two (both have ultrasonic sensors).

I would not be so sure if I were you. Disabling the existing USS is just an “update” away. The question is what will come first - disabling of the existing USS or releasing something using the almighty TV.

 

[Battpower]

Not worth my time. Stopped buying Teslas in 2018 so it doesn't affect either of my current two (both have ultrasonic sensors). As it stands now, I have no plans to buy more, for one I cannot live with stalkless steering, but also would want parking sensors and learned over the years that if it's not available on delivery day, good chance it will never be available for that car - you might get an excuse from Tesla, like "it was only a forward looking aspiration goal" not "a promise of any specific functionality".

Based on the fairly aligned ouija boards of those on this thread, you might want to start willing Tesla to fail in their attempts at dispensing with the need for USS if you want to keep your USS functional. Mind you, you must be holding off on updates anyway if you are keeping radar functional.

So it seems like you might be stuck between rock and hard place - you get to keep USS (and radar may be) or you and your friend get to test out Tesla's next VO project.

 

[whitex]

Quick update. I had my relative ask Tesla sales if they would delay the car delivery from next week to until the camera based parking system is working, since they said it's coming really soon (it's been 4 months already since it's been coming soon, right?). They laughed and said no, they are cannot hang onto the car (and guarantee its price) for that long.

 

[LiamPope]

Quick update. I had my relative ask Tesla sales if they would delay the car delivery from next week to until the camera based parking system is working, since they said it's coming really soon (it's been 4 months already since it's been coming soon, right?). They laughed and said no, they are cannot hang onto the car (and guarantee its price) for that long.

I would say that suggests they know something and are implying it's not coming any time soon.

But I know Tesla sales staff know nothing at all.

 

[SalisburySam]

I would not be so sure if I were you. Disabling the existing USS is just an “update” away. The question is what will come first - disabling of the existing USS or releasing something using the almighty TV.

Oh, I think we all know the answer to that, based on the treatment given vehicles with radar. DisableVision still hasn’t reached functional parity with radar-equipped cars now many months later. Why would USS disabling be different?

 

[AlanSubie4Life]

Tesla has no plans to disable USS and is unlikely to change those plans until the substitute is really good and better than USS in all cases (which will be effectively never).

Plus, unlike with radar, there is no downside and essentially zero overhead to just leaving the USS enabled.

There is no reason to disable them, and it is *more work* for Tesla to do so - they would have to disable them (so again very unlikely until the TeslaVision version is spectacular and perfect).

 

[emmz0r]

Quick update. I had my relative ask Tesla sales if they would delay the car delivery from next week to until the camera based parking system is working, since they said it's coming really soon (it's been 4 months already since it's been coming soon, right?). They laughed and said no, they are cannot hang onto the car (and guarantee its price) for that long.

will there be a class action like the AP2 when they didnt reach feature parity?

 

[Boza]

will there be a class action like the AP2 when they didnt reach feature parity?

I seriously doubt it. “Feature parity” would be very difficult to prove and argue.
BTW: Expect the same to happen with FSD. Legacy cars will receive something that “checks the box” and not necessarily match people’s expectations from the time they bought the car. Which emphasizes the rule that you should never buy a product based on promised future functionality.

 

[Boza]

Tesla has no plans to disable USS and is unlikely to change those plans until the substitute is really good and better than USS in all cases (which will be effectively never).

Plus, unlike with radar, there is no downside and essentially zero overhead to just leaving the USS enabled.

There is no reason to disable them, and it is *more work* for Tesla to do so - they would have to disable them (so again very unlikely until the TeslaVision version is spectacular and perfect).

This has been discussed many times but here is a recap of the incentives behind disabling the existing USS:
- Single code base. Less development and simplified testing -> lower cost.
- Simplified inventory management (similar to the above but for hardware). It looks like it is not as simple as drilling holes; there is supporting structure that is molded into the bumper. So, removing the USS simplifies the inventory they need for maintenance - keep only non-USS bumpers and if USS car needs a bumper replacement just update the firmware and slap a non-USS bumper. That simplifies the technicians training as well. Even if it is as simple as drilling holes that would mean you have to have the necessary equipment and training to support it. (Side note: folks, be very careful about accidents if you want to keep your USS)
- Implementation of the Tesla Vision dogma. This is a religious argument so it may be more powerful than the other two in terms of accelerating the disablement.

 

[enemji]

Funny you mentioned that. When auto wipers was first released, that's exactly what Tesla said: that it would use driver interventions of the wipers to improve the AI. Not really sure if they ever followed through and for how long. But there's a big flaw with this approach: if the wipers go on when it shouldn't, there's no way for the driver to provide an interventional feedback to tell it not to wipe.

sorry, OT. carry on.

They have the data. It is my presumption and understanding that the wipers not working correctly and users intervening must lie in the tiny statistical error zone and hence not acceptable for corrective actions.

We have 2xM3LR for 12 months now. We have driven them in all kinds of weather permutations and combinations of (Sunny, cloudy, dark nights) x (drizzling, regular rains, heavy thunderstorms) (except snow/ice) and using autosteer in those conditions. No issues with the wipers whatsoever.

 

[LiamPope]

This has been discussed many times but here is a recap of the incentives behind disabling the existing USS:
- Single code base. Less development and simplified testing -> lower cost.
- Simplified inventory management (similar to the above but for hardware). It looks like it is not as simple as drilling holes; there is supporting structure that is molded into the bumper. So, removing the USS simplifies the inventory they need for maintenance - keep only non-USS bumpers and if USS car needs a bumper replacement just update the firmware and slap a non-USS bumper. That simplifies the technicians training as well. Even if it is as simple as drilling holes that would mean you have to have the necessary equipment and training to support it. (Side note: folks, be very careful about accidents if you want to keep your USS)
- Implementation of the Tesla Vision dogma. This is a religious argument so it may be more powerful than the other two in terms of accelerating the disablement.

For simple manual park assist and low speed proximity sensing - the code is USS only and finished and working. No maintenance overhead if they simply leave that in and at most put a toggle in the menu to toggle between the old USS version and the new camera version when it comes.

I agree for summon and autopark which are still in development and use USS and cameras that there would be an overhead there - they'd have to keep updating both a camera only version and USS plus camera, and they won't do that. But just keep USS park assist in there for the love of god and keep the USS working. For summon and autopark just don't use them.

 

[Boza]

For simple manual park assist and low speed proximity sensing - the code is USS only and finished and working. No maintenance overhead if they simply leave that in and at most put a toggle in the menu to toggle between the old USS version and the new camera version when it comes.

I am not familiar with the code base but would suspect that it is integrated somewhere (I.e. regression testing needed) and must be maintained, as everything else (I.e. at least continuous automated testing supported by someone).
There is a joke in the dev community that MS still maintains the mouse pointer code. If there is code then it needs maintenance, however little it may be.

 

[ChaoticNeutral]

Don’t think USS are used for AP.

They definitely are, otherwise “auto lane change unavailable cannot communicate with USS” and “navigate with autopilot unavailable due to USS communication issue” alerts wouldn’t show up I imagine…

 

[whitex]

will there be a class action like the AP2 when they didnt reach feature parity?

6 years from now I guess.

 

[whitex]

I would say that suggests they know something and are implying it's not coming any time soon.

But I know Tesla sales staff know nothing at all.

They obviously know that "soon" and "near future" could mean many years to never.

 

[Mullermn]

They definitely are, otherwise “auto lane change unavailable cannot communicate with USS” and “navigate with autopilot unavailable due to USS communication issue” alerts wouldn’t show up I imagine…

My car does not have USS and I can use AP and NoA fine. Maybe they’re used if you have them but they cannot be essential.

Have you seen those messages recently? From a liability angle I’d be surprised if Tesla rolled out code for a safety function like AP with two modes like that. Either the sensors are needed or they’re not, plus that’s two code variants to test and we know they’re not keen on duplication.

 

[ChaoticNeutral]

My car does not have USS and I can use AP and NoA fine. Maybe they’re used if you have them but they cannot be essential.

Have you seen those messages recently? From a liability angle I’d be surprised if Tesla rolled out code for a safety function like AP with two modes like that. Either the sensors are needed or they’re not, plus that’s two code variants to test and we know they’re not keen on duplication.

Yes, I see them and “parking assist unavailable” - one of my USS got struck by another vehicle, AP auto lane change and NOA no longer work and I get these errors, summon also doesn’t work. There’s not much other obvious damage besides paint scuffs near where my vehicle was hit.

FSD is fully functional on streets, including lane changes FWIW. Those messages still show, but it’s a different stack. I did just get an update alert about 10 minutes ago so maybe that’ll get rid of the USS need for NOA and Summon?

 

[jeewee3000]

Hi all, just wanted to post an update. I asked my local Tesla dealership (Belgium) when the Tesla Vision update for no-USS cars is to be expected.
Their reply (deepl translation):

"Unfortunately, we have no idea when to expect this update.
We hope together with our customers that this will happen soon but cannot make any promises in this regard."

So Tesla representatives/employees outside the AI Team generally have no clue and probably know less than we do.
I'm cross-posting this in the three main "no-USS threads".

 

[whitex]

I think Tesla should simply subtract a price of the parking sensors retrofit from each new car being sold, then charge for the software once it's available. It seems like a fair solution, and gives Tesla great incentive to get it done. Without such a deal, Tesla already has customers' money, there is no incentive whatsoever to complete this - if there is a choice to divert Tesla engineers to fix some latest Twitter problem or work on parking sensor software, guess which one is better for Elon's bottom line.

 

[Boza]

That is pretty much how other manufacturers dealt with the USS shortage.