2023 Holiday Update

[yessuz]

As I understand it these S3XY things are injecting CAN signals the same way the OEM controls do, so I don’t know how the car would even “know” anything foreign was doing it.

If I were a Tesla coder tasked with trying to block this stuff I’d be looking at how precise the “reactivate AP” signal was, and if it was happening within the same timeframe (I presume this does it precisely) then it could be detected as automated, and therefore illegitimate. The same sort of thing is done for wheel weights, requiring variable torque etc.

So code the delay to re-enable AP within timeframe at random. Let's say you chose between 2-4 s and randomize.

All is possible

 

[Durzel]

So code the delay to re-enable AP within timeframe at random. Let's say you chose between 2-4 s and randomize.

All is possible

Yeah, it would be whack-a-mole if Tesla tried to stop it. Easy enough to stop a fixed weight wheel thing, not to easy to block something that can be updated.

 

[BitJam]

Do not kid yourself and others about FSD. It is not there and never be in UK. Not with this attitude

I said Tesla was investing billions of dollars into FSD and was devoting most of the software resources to it, to the detriment of software improvements in other areas. How did you make the leap from that statement to me kidding myself and others about the state of FSD?

FSD 11 is unsafe and doesn't work where I live in rural New Mexico. It can't even get me to the grocery store 5 minutes away. By stressing the large investment in money and resources Tesla is pouring into FSD to get it to do what it says on the tin, I didn't mean to imply that goal was near at hand. Normally people pull resources away from projects that are winding down while they dump vast amounts of resources into projects that still need a lot more work.

OTOH I do expect autonomous driving to reach all of the US and Europe eventually although I have no idea when this will happen. My guess is 2 to 10 years. It is unclear to me what attitude you think is going to prevent this from ever happening.

But instead of adding goat scream for car lock or another fart sound, they should just spend 1 hr to improve AP.

The software talent needed to add goat scream and fart noise window dressing is miles away from the talent needed to make significant changes to AP. If you are unfamiliar with why software engineers are not interchangeable, I recommend the Mythical Man Month (which is now almost 50 years old).

But more importantly, Tesla has yet again totally revamped their FSD architecture, dropping 300K lines of human written code. For over 6 months I have been saying this is why they are not making "obvious" improvements to V11 AP and FSD. All of that work will get thrown away (along with the other 300K lines of code) when they switch to V12. This is part of the reason I said:

My perspective is that Tesla is putting billions of dollars and most of the[ir] software development effort into getting FSD to actual do what it says on the tin. This means that a lot of software improvements, especially those that improve the manual or semi-manual driving experience, are lower priorities.​

It seems obvious that having Tesla divert scarce resources away from FSD to add minor features to AP that are going to get thrown onto the scrap heap in less than a year is not the path to solving FSD as quickly as possible. OTOH, if Tesla does adopt this strategy then your prediction of them never solving FSD may be correct.

 

[Drew57]

My perspective is that Tesla is putting billions of dollars and most of the[ir] software development effort into getting FSD to actual do what it says on the tin. This means that a lot of software improvements, especially those that improve the manual or semi-manual driving experience, are lower priorities. It seems obvious that having Tesla divert scarce resources away from FSD to add minor features to AP that are going to get thrown onto the scrap heap in less than a year is not the path to solving FSD as quickly as possible.​

Perhaps I'm in the minority but a manual or semi manual driving experience is all I want and if they are lower priority then this might as well be my last Tesla. That would be a great shame as I genuinely like these cars, mostly for those very aspects that you (maybe correctly) infer will be phased out.

Some of us like to interact with things & enjoy being part of the experience rather than just passively watching as it happens around us. However if I were to own the Self Driving car as imagined by Tesla I might as well use taxis and save a lot of money into the bargain..... but where's the fun in that?

It also seems something of a paradox that a manufacturer that has a goal to take the human element out of driving continues to develop some of the fastest accelerating road vehicles on the planet. Safe, steady and autonomous as the No1 priority combined with supercar performance but ultimately no need for a driver to enjoy that.

 

[M1tch]

If only Tesla could link the internal camera to the windscreen wipers and we’d have a half decent system!

What, so when it lip reads “oh FFS just wipe the window” it does?

 

[yessuz]

I said Tesla was investing billions of dollars into FSD and was devoting most of the software resources to it, to the detriment of software improvements in other areas. How did you make the leap from that statement to me kidding myself and others about the state of FSD?

FSD 11 is unsafe and doesn't work where I live in rural New Mexico. It can't even get me to the grocery store 5 minutes away. By stressing the large investment in money and resources Tesla is pouring into FSD to get it to do what it says on the tin, I didn't mean to imply that goal was near at hand. Normally people pull resources away from projects that are winding down while they dump vast amounts of resources into projects that still need a lot more work.

OTOH I do expect autonomous driving to reach all of the US and Europe eventually although I have no idea when this will happen. My guess is 2 to 10 years. It is unclear to me what attitude you think is going to prevent this from ever happening.



The software talent needed to add goat scream and fart noise window dressing is miles away from the talent needed to make significant changes to AP. If you are unfamiliar with why software engineers are not interchangeable, I recommend the Mythical Man Month (which is now almost 50 years old).

But more importantly, Tesla has yet again totally revamped their FSD architecture, dropping 300K lines of human written code. For over 6 months I have been saying this is why they are not making "obvious" improvements to V11 AP and FSD. All of that work will get thrown away (along with the other 300K lines of code) when they switch to V12. This is part of the reason I said:

My perspective is that Tesla is putting billions of dollars and most of the[ir] software development effort into getting FSD to actual do what it says on the tin. This means that a lot of software improvements, especially those that improve the manual or semi-manual driving experience, are lower priorities.​

It seems obvious that having Tesla divert scarce resources away from FSD to add minor features to AP that are going to get thrown onto the scrap heap in less than a year is not the path to solving FSD as quickly as possible. OTOH, if Tesla does adopt this strategy then your prediction of them never solving FSD may be correct.

Seriously!?

It took some third party Bulgarian engineers to write some code like:
IF AP is enabled AND AP turned Off due to lane change THEN Enable AP after Lane Change.

I mean even for me it looks quite straightforward. And if there is no coding Talent in Tesla to enable such simple life improvement but we celebrate goat or custom sound, so where is that talent for FSD and wtf are they doing? Wtf is tesla doing at all? All manufacturers already overtook AP in this regard and we still pretend that we are latest and greatest!?

Seriously you cannot and should not defend this stupidity.

also, FSD is "not reliable in New Mexico vs FSD Not Available in rest of the world are two quite different things

 

[hmadsen]

As I understand it these S3XY things are injecting CAN signals the same way the OEM controls do, so I don’t know how the car would even “know” anything foreign was doing it.

If I were a Tesla coder tasked with trying to block this stuff I’d be looking at how precise the “reactivate AP” signal was, and if it was happening within the same timeframe (I presume this does it precisely) then it could be detected as automated, and therefore illegitimate. The same sort of thing is done for wheel weights, requiring variable torque etc.

They can probably also just change it so that the CAN signal sent and received is suddenly changed.

Yes, S3xy buttons can change something too, but then it will be a matter of pingping, where it only works "half of the time".

 

[WannabeOwner]

S3XY Buttons provide this feature. They call it "continuous autopilot".

Bjorn demo'd this (or something like it) in a recent video. I presume they were S3XY buttons (but I wasn't playing close attention)

EDIT: Re-Checked the video @ 4:00, definitely S3XY buttons, and he demo's a use for engaging AP and the S3XY lane-change at 5:20 - 7:15

Its part of a LR Highland range test in Winter (at minus 20C !!)

 

[hmadsen]

Bjorn demo'd this (or something like it) in a recent video. I presume they were S3XY buttons (but I wasn't playing close attention)

EDIT: Re-Checked the video @ 4:00, definitely S3XY buttons, and he demo's a use for engaging AP and the S3XY lane-change at 5:20 - 7:15

Its part of a LR Highland range test in Winter (at minus 20C !!)

Is this just re-engaging after a fixed amount of seconds ?

I once engaged it too soon while i had basically not completed the lanechange and the car freaked out because it did not know what lane it should select.

I would imagine if someone is a bit slow in changing lanes they would run the risk of encountering this if it just re-enables after it had been disabled for lane change after a few seconds.

Will it always re-engage in this configuration ? What if i manually disables it for example if i am approaching a roundabout, will it then just auto-enable a few seconds later or is it clever enough only to re-engage if i am on a multilane street and have it turned on and disengage it with steering wheel input and blinker ?

Potentially it could enable the AP at some point where i dont really want it to ?

 

[KennethS]

Is this just re-engaging after a fixed amount of seconds ?

I once engaged it too soon while i had basically not completed the lanechange and the car freaked out because it did not know what lane it should select.

I would imagine if someone is a bit slow in changing lanes they would run the risk of encountering this if it just re-enables after it had been disabled for lane change after a few seconds.

Will it always re-engage in this configuration ? What if i manually disables it for example if i am approaching a roundabout, will it then just auto-enable a few seconds later or is it clever enough only to re-engage if i am on a multilane street and have it turned on and disengage it with steering wheel input and blinker ?

Potentially it could enable the AP at some point where i dont really want it to ?

It reengages within a chosen time after the indicator is turned off following the completion of the lane change. So unless you manually cancel the indicator before completing the lane change your (valid) concern won't be an issue. As to your second question, the function won't operate after a manual disengagement of AP. It only gets triggered when AP shuts off as a result of turning on an indicator. I don't use the function since we have FSD and/or EAP but that is my understanding having read and participated in a Enhance group on Discord. People who have it and use it seem very happy with it.

 

[hmadsen]

It reengages within a chosen time after the indicator is turned off following the completion of the lane change. So unless you manually cancel the indicator before completing the lane change your (valid) concern won't be an issue. As to your second question, the function won't operate after a manual disengagement of AP. It only gets triggered when AP shuts off as a result of turning on an indicator. I don't use the function since we have FSD and/or EAP but that is my understanding having read and participated in a Enhance group on Discord. People who have it and use it seem very happy with it.

Well this seems to be a much cheaper alternative to buying EAP or FSD.

To be honest i would much rather just buy the EAP package (FSD makes not sense here as the only 2 things gained over EAP is autostop at red light and STOP signs) but for some strange reason Tesla still wants silly money for that function.

Here the EAP package cost about $4300 but i would argue if they slashed the price to $1000 they would probably sell more than 4 times as many of these.

This price is by the way the same as they asked in 2022 when the price of a Tesla was $26.000 higher.

I would argue it would be logical to lower the price of the packages when they lower the price of the car.

I am going to look into getting these s3xy buttons or the s3xy bar, i assume this function is actually a function managed by the commmander though.

My only concern is...If this becomes a very popular alternative to buying the EAP package, i would still argue, there is a risk, Tesla would do something to mitigate this, if they think it would hurt their sales of EAP/FSD

 

[Pagemakers]

My only concern is...If this becomes a very popular alternative to buying the EAP package, i would still argue, there is a risk, Tesla would do something to mitigate this, if they think it would hurt their sales of EAP/FSD

Financially it’s not a large risk

 

[hmadsen]

Financially it’s not a large risk

I agree, but in most cases in life it is easier to never have had a feature, rather than having had it and lost it.

There is not a lot the people behind S3Xy buttons can do, if Tesla decides they dont want them to be able to do what they do, they can change something in every update to "break" the integration, and if they persist, they risk Tesla doing something to make S3XY buttons completely useless.

It seems the buttons do their "commands" via the phone and bluetooth, but will this function also work just with the buttons or do i need the commander also ?

 

[KennethS]

Well this seems to be a much cheaper alternative to buying EAP or FSD.

To be honest i would much rather just buy the EAP package (FSD makes not sense here as the only 2 things gained over EAP is autostop at red light and STOP signs) but for some strange reason Tesla still wants silly money for that function.

Here the EAP package cost about $4300 but i would argue if they slashed the price to $1000 they would probably sell more than 4 times as many of these.

This price is by the way the same as they asked in 2022 when the price of a Tesla was $26.000 higher.

I would argue it would be logical to lower the price of the packages when they lower the price of the car.

I am going to look into getting these s3xy buttons or the s3xy bar, i assume this function is actually a function managed by the commmander though.

My only concern is...If this becomes a very popular alternative to buying the EAP package, i would still argue, there is a risk, Tesla would do something to mitigate this, if they think it would hurt their sales of EAP/FSD

I have to agree that EAP is overpriced and FSD pricing is unjustifiable. When we purchased our 2019 M3P with FSD the choice was only FSD or nothing (and we knew far less back then although the price was lower than now). With our Y EAP was a choice and we went with it knowing it has flaws and is overpriced but does include features we find useful.

 

[KennethS]

I agree, but in most cases in life it is easier to never have had a feature, rather than having had it and lost it.

There is not a lot the people behind S3Xy buttons can do, if Tesla decides they dont want them to be able to do what they do, they can change something in every update to "break" the integration, and if they persist, they risk Tesla doing something to make S3XY buttons completely useless.

It seems the buttons do their "commands" via the phone and bluetooth, but will this function also work just with the buttons or do i need the commander also ?

I believe the commander is a requirement. We find the buttons useful but I understand they will not appeal to most people.

 

[WllXM]

I have to agree that EAP is overpriced and FSD pricing is unjustifiable. When we purchased our 2019 M3P with FSD the choice was only FSD or nothing (and we knew far less back then although the price was lower than now). With our Y EAP was a choice and we went with it knowing it has flaws and is overpriced but does include features we find useful.

I would say EAP without USS today is indeed completely overpriced.
If they bring back Summon and Autopark, it may still be expensive, but a bit more in line with Driving Assist optional extras from other automakers.

 

[KennethS]

I would say EAP without USS today is indeed completely overpriced.
If they bring back Summon and Autopark, it may still be expensive, but a bit more in line with Driving Assist optional extras from other automakers.

I have not been in a Tesla without USS but from what I read here it seems you are right.

 

[hmadsen]

I would say EAP without USS today is indeed completely overpriced.
If they bring back Summon and Autopark, it may still be expensive, but a bit more in line with Driving Assist optional extras from other automakers.

Yes this is actually the most "infuriating" thing.

They are basically asking full price for a package that only has 2 of 5 features available and a vague "promise" that the other 3 features will eventually re-appear.

So

Pay full price for 40% of the features and a "trust me bro" for the rest

 

[bubblysun]

I use a lot of autopilot in UK and am still undecided whether to update from 2023.38.

I have a model Y with great parking sensors and nothing in the update is more concerning than the instant autopilot nags… Everything just ‘works’ at the moment.

Should I update to 2023.44?

 

[kelvin 660]

I'm on 2023.44.30.9 and I haven't seen a big difference between 2023.38, apart from when it nags it will display a warning message, whereas before I'd get the blue sky and sometimes a beep. I haven't had a yellow card, yet.
By the way, my car insurance (Churchill) states that the car should always be running the latest software.