Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

v11 software update SUCKS

Do you prefer v11 Tesla UI to v10.x, or want Tesla to go back to v10?


  • Total voters
    591
This site may earn commission on affiliate links.
So we're stuck with an excellent piece of EV machinery 5-6 generations above anything the competition has to offer, but with a brain dead interface.
Time for an opensource aftermarket MCU that can plug and go. Maybe one that doesn't support autopilot and all that baloney, but one that is customizable, snap in, and geared for those people who still want to DRIVE the car. - P3 and Plaid owners!
 
Madness. My 3.5 years love affair with my Model 3 ended on December 25th when Elon & co entered my $80k car and removed buttons that I relied on for safety and ease of use.
I feel a similar disappointment - is there nothing in a legal vein that prevents Tesla for forcing cosmetic changes on us? Most especially when anyone can make a strong argument that safety is compromised by some changes?

It's my damn car after all
 
I think what Elon meant was that if a human needs to take over from the AI, the AI has failed in it's job and that becomes an area where the AI needs to be improved. I did not interpret it as a criticism on the human drivers.
I believe this is exactly what he meant. If you listen to this video the context in which he made this statement is very clear. Elon is discussing the AP attempting to solve for the best path to take through a complex intersection. He explains that if the human driver had to intervene then this action is effectively an error indication to the network. If we assume the human driver is the correct response, for example, then technically the error is the difference between what the autopilot wanted to do and what the human wanted to do.

1:14, "The ones where there are no interventions, are the right ones".

 
Last edited:
  • Disagree
Reactions: ContrarianDC
I believe this is exactly what he meant. If you listen to this video at 1:56 the context in which he made this statement is very clear. Elon is discussing the AP attempting to solve for the best path to take through a complex intersection. He explains that if the human driver had to intervene then this action is effectively an error indication to the network. If we assume the human driver is the correct response, for example, then technically the error is the difference between what the autopilot wanted to do and what the human wanted to do.

But in the tweet that was posted in this thread, he made that statement in reference to a mock up someone made with climate controls placed back on the bottom of the main screen. That seems to indicate that the car should know better than us how warm or cool we prefer our bodies to be. I have tried auto seat warmers since the v11 update. It doesn’t work because sometimes I want to feel warmer or cooler. That is always going to be personal preference and not something that the car can choose for us.
 
But in the tweet that was posted in this thread, he made that statement in reference to a mock up someone made with climate controls placed back on the bottom of the main screen. That seems to indicate that the car should know better than us how warm or cool we prefer our bodies to be. I have tried auto seat warmers since the v11 update. It doesn’t work because sometimes I want to feel warmer or cooler. That is always going to be personal preference and not something that the car can choose for us.
I know, and it is a mystery to me why he reused this quote to address that UI video. My best guess so far is he just responded quickly without giving it much thought. I'm still trying to figure that one out.
 
I see it as buying time until viable alternatives appear. I love the car but the trend is concerning and, if they continue this way, that will be my last Tesla :(

I feel this same way. I've believed that the M3 is the world's best car and gave everyone I could a test drive (strangers even). Now I wouldn't recommend it... and I would've felt *really bad* if someone bought one on my recommendation and then ended up with v11.
 
  • Like
Reactions: Msjulie
So we're stuck with an excellent piece of EV machinery 5-6 generations above anything the competition has to offer, but with a brain dead interface.
Time for an opensource aftermarket MCU that can plug and go. Maybe one that doesn't support autopilot and all that baloney, but one that is customizable, snap in, and geared for those people who still want to DRIVE the car. - P3 and Plaid owners!
they could do this and make it safe (from an api POV but also functional safety) - but they'd have to invest a lot of their staff to this and that might cause some farts or light shows to be delayed for next year.

sigh.

they wont open it for lots of reasons, some of which I actually agree with. but it could be done. it would need some serious red team work to ensure there are no (obvious, at least) sec holes and I dont think they'd put the right staff to that (if they could even get it; they are not the shining star of vehicle security..)

the idea of a snap-in user interface, either physically or in code, that should come and the first vendor to safely architect this will get my next car money. but its a hard sell and so far, the car companies dont seem to show any interest in allowing users to really customize things to that level.
 
  • Like
Reactions: Hank42
I know, and it is a mystery to me why he reused this quote to address that UI video. My best guess so far is he just responded quickly without giving it much thought. I'm still trying to figure that one out.

Or he's really focused on FSD at the moment, so when the topic of "need better UI for human input" came up, his mind associated the two and he burted a response.
This is probably the best answer. He might not have even watched the video.

However, his answer tracks with the push to automate all the things in the video (climate, defrost, and anything else).

For us, ALL the auto anything is exactly broken. Nothing works AT ALL when set to Auto. Lights, climate, wipers, lane keeping, I even had to turn off emergency lane assist and lane warnings driving I90 from Seattle to Spokane 4 days ago: snow, rain, packed ice on the road rendered everything useless, and twice the E lane assist tried to drive me off the actual lane into snow berms.

New philosophy: Almost all human input is in response to system settings that are wrong right now (seat too hot/cold, cabin too hot/cold, wipers on too high/low/off, navigating to wrong supercharger (predicted SOC too high/low).

There is no time I want to change something except when the system is wrong. Give me easy access to anything that could be wrong.

Dead Horse=beat, I know, but this is an intensely strange situation.
 
So we're stuck with an excellent piece of EV machinery 5-6 generations above anything the competition has to offer, but with a brain dead interface.
Time for an opensource aftermarket MCU that can plug and go. Maybe one that doesn't support autopilot and all that baloney, but one that is customizable, snap in, and geared for those people who still want to DRIVE the car. - P3 and Plaid owners!
Totally replacing the MCU would be very challenging. Not sure you could get supercharging working with that.
 
  • Disagree
Reactions: Silicon Desert
Open API UI sounds like a good idea but it is very dangerous because there is no air gap between the entertainment system and car controls. Open source QA is even more volatile than Tesla - you could get something excellent or something, that looks good, but disables your heat pump (pun intended). Also, there is the issue of liability.
Good idea but not possible unless there is an air gap between critical and entertainment functionality.
 
they could do this and make it safe (from an api POV but also functional safety) - but they'd have to invest a lot of their staff to this and that might cause some farts or light shows to be delayed for next year.

sigh.

they wont open it for lots of reasons, some of which I actually agree with. but it could be done. it would need some serious red team work to ensure there are no (obvious, at least) sec holes and I dont think they'd put the right staff to that (if they could even get it; they are not the shining star of vehicle security..)

the idea of a snap-in user interface, either physically or in code, that should come and the first vendor to safely architect this will get my next car money. but its a hard sell and so far, the car companies dont seem to show any interest in allowing users to really customize things to that level.
I think @wk057 mentioned writing replacement software, but it was a passing comment ages ago that I'm not even sure was serious. I only remember it because it gave me a little hope.
The things you mention, linux-works, don't make money. They're good ideas, but they don't sell cars. The games, light shows, farts, and all that stuff get people talking. They make a memorable impression on newcomers. They make the news. That's not for owners. It's advertising. That's the priority.
 
Totally replacing the MCU would be very challenging. Not sure you could get supercharging working with that.
Yes, it would be challenging. But so was climbing Mt. Everest, but as a species, we did it! Is this worth doing - probably not, but I figured I'd pull the thread and start the conversation at least. Heck, even leaving the existing MCU in place and hopping in the CAN bus with an alternate interface might be worth doing for dedicated controls.
 
Yes, it would be challenging. But so was climbing Mt. Everest, but as a species, we did it! Is this worth doing - probably not, but I figured I'd pull the thread and start the conversation at least. Heck, even leaving the existing MCU in place and hopping in the CAN bus with an alternate interface might be worth doing for dedicated controls.
Supplemental controls for some functions are absolutely possible. There are already some on the market (S3XY buttons, etc), but they all seem to have their limitations. Certainly a lot more room for innovation in this space, and looking like a good market to get into.
 
But in the tweet that was posted in this thread, he made that statement in reference to a mock up someone made with climate controls placed back on the bottom of the main screen. That seems to indicate that the car should know better than us how warm or cool we prefer our bodies to be. I have tried auto seat warmers since the v11 update. It doesn’t work because sometimes I want to feel warmer or cooler. That is always going to be personal preference and not something that the car can choose for us.
I interpreted his tweet to mean that he didn’t think third parties should be messing with the interface. In general I agree, but I’d also say 3rd parties shouldn’t need to mess with the interface!
Open API UI sounds like a good idea but it is very dangerous because there is no air gap between the entertainment system and car controls. Open source QA is even more volatile than Tesla - you could get something excellent or something, that looks good, but disables your heat pump (pun intended). Also, there is the issue of liability.
Good idea but not possible unless there is an air gap between critical and entertainment functionality.
No, I would strongly argue against an open API UI. Think about the potential problems that opens up. Right now, Tesla has access to everything, sole control and the can easily test the code they write. If they open up the code they need to change the configuration of the OS so it has public and private APIs, they’d need to develop and maintain programmer resources, they’d need to develop a protocol for installing 3rd party apps, they’d have to worry about security and sandboxing and they’d have to worry about bugs in the software taking down the system. Anyone who’s installed software on their computer with a bug or that’s interfered with other software or caused it to crash can appreciate this.

I don’t like V11 but I think Tesla should fix it and bringing 3rd parties in will only complicate the picture.
 
they could do this and make it safe (from an api POV but also functional safety) - but they'd have to invest a lot of their staff to this and that might cause some farts or light shows to be delayed for next year.

sigh.

they wont open it for lots of reasons, some of which I actually agree with. but it could be done. it would need some serious red team work to ensure there are no (obvious, at least) sec holes and I dont think they'd put the right staff to that (if they could even get it; they are not the shining star of vehicle security..)

the idea of a snap-in user interface, either physically or in code, that should come and the first vendor to safely architect this will get my next car money. but its a hard sell and so far, the car companies dont seem to show any interest in allowing users to really customize things to that level.
They have the capital to do it and to do it right. This is a failure of the leadership to not just prioritize correctly but to go to the board and get new teams stood up. Eg why aren’t they crowd sourcing map fixes? They are in a unique position to automate that and it could be a strato advantage. Or ‘we need to fund & hire a usability person/team’. QA team.