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

Are valuable FSD features coming soon suggesting FSD purchase now?

This site may earn commission on affiliate links.
No, just look at EAP development. When they were releasing Navigate on Autopilot they decided to keep lane change confirmation mandatory. They wanted another 10M miles first

Given the size of the fleet 10 million AP miles is (depending on your estimates) anywhere from 1-2 to perhaps 3-4 weeks of time.
(~150k AP2.x Teslas exist in the US, say half have EAP (it was 77% last report I saw on the S/X fleet, let's say it's WAY lower on 3 so 50% of fleet)... that gets you 75k cars. Average driver in the US does ~30 miles a day. Let's say only half those are using EAP (probably low- as those using it most likely drive more than average)... So that's 75,000 cars at 15 miles a day- or 1.125 million miles of EAP data. Per day. Less than 9 days to get 10 million miles. And that's making pretty conservative estimates. Even if it's HALF those conservative guesses, you're still talking less than a month to get 10 million miles of data.


They barely started developing the enhanced part of EAP.

Huh?

Other than summon going around corners, EAP is completely done with Drive on Nav (on ramp to off ramp)

Obviously the existing features will continue to get better, but there's no more features coming after that. Anything else will be FSD.

And the first ones as soon as the HW3 is deployed supposedly.
 
  • Funny
Reactions: P85_DA and outie
Yeah, it was pretty crappy. They had some trouble loading the software after the hardware changes were made, which caused them to call up headquarters to sort it out. That is when the local service center folks were told to cease and desist - it was no longer supported. Couldn't MAKE them upgrade our car, and it wasn't like they were trying to charge us $$ for services not performed, so we dropped it. They at least comped us our yearly maintenance that we were also getting done, so there was that. Would rather have gotten the L tho.

When you provided details it makes a little more sense. Not that it doesn’t suck it does but does provide some context.

Your L software depended on talking to the mothership. If the mothership won’t let the software download, nothing the center can do.

I have a unicorn X in the sense that it has bio defense but cloth seats. I added bio as an upgrade option. Lots of people were saying they could save thousands by skipping PUP and upgrade ala carte. Option lasted for one week only I believe.

I also added tow kit post purchase. I even warned people here that if they wanted it ever.. do it before the Model 3s arrive.

It’s been “out of stock” for 18 months. How can it be out of stock when every new X gets it. It’s out of stock because no money that makes sense to us makes sense to Tesla to retrofit you.

So yeah skip FSD pre order at your peril.
 
  • Like
Reactions: Jedi2155 and P85_DA
There are plenty of videos showing that nav on autopilot will kill you if you’re not paying attention.


So will regular autopilot. Or just normal driving.

Paying attention while driving is kind of important in any car not offering at least L3 features (which is none of them other than the Audi in europe in limited circumstances.

If that'll change at all with HW3/FSD remains to be seen... but EAP will based on anything we've seen remain an L2 system requiring driver attention.

I don’t believe that they have a software build that fixes that. I also don’t believe that the primary problem is the speed of the processor. I’m not an expert though.

What do you think it is then?
 
There are plenty of videos showing that nav on autopilot will kill you if you’re not paying attention. I don’t believe that they have a software build that fixes that. I also don’t believe that the primary problem is the speed of the processor. I’m not an expert though.

I’m not an expert either so I won’t say I’m speaking as one. :)

AI can win through brute force. Those chess appliances can crush any human chess master now because appliances got to the point where they can just calculate every possible move and react accordingly.
 
I could see Tesla releasing some minor FSD features before HW3 is installed in vehicles, they'd probably be features that are less computationally expensive. Perhaps something like speed limit sign recognition (don't think that is EAP domain...). I'd love to see some features come soon to help those of us who are on the fence regarding purchasing the package.

I agree with those who say that Tesla is under no obligation to offer the FSD package on into infinity, and it seems Elon alluded to limiting the availability in the last quarterly call (maybe I heard wrong).

I'm an FSD bear, but also recognize the tea leaves that seem to indicate features are on the horizon and Tesla has plans to change the structure of the FSD package.

One drum I want to beat is regarding the hardware (not to be confused with the FSD software package). Tesla has sold and is still selling vehicles advertised as having all needed hardware for FSD. So folks need to stop worrying about needing to purchase FSD in order to avoid an up-charge for HW3. Tesla cannot and will not charge you for the hardware upgrade. You paid for a vehicle that has all required FSD hardware and if FSD requires a retrofit Tesla has to provide that at no additional charge. Now, if HW3 is of no benefit to EAP then peace, there is arguably no obligation on Tesla's part to upgrade your computer. But Tesla isn't going to charge a HW2 car more for FSD than a HW3 car.
 
Given the size of the fleet 10 million AP miles is (depending on your estimates) anywhere from 1-2 to perhaps 3-4 weeks of time.
(~150k AP2.x Teslas exist in the US, say half have EAP (it was 77% last report I saw on the S/X fleet, let's say it's WAY lower on 3 so 50% of fleet)... that gets you 75k cars. Average driver in the US does ~30 miles a day. Let's say only half those are using EAP (probably low- as those using it most likely drive more than average)... So that's 75,000 cars at 15 miles a day- or 1.125 million miles of EAP data. Per day. Less than 9 days to get 10 million miles. And that's making pretty conservative estimates. Even if it's HALF those conservative guesses, you're still talking less than a month to get 10 million miles of data.




Huh?

Other than summon going around corners, EAP is completely done with Drive on Nav (on ramp to off ramp)

Obviously the existing features will continue to get better, but there's no more features coming after that. Anything else will be FSD.

And the first ones as soon as the HW3 is deployed supposedly.
Just because EM said 10M does not mean it will happen at 10M. Please, we’ve been there.

They’ve said AP2 was capable of FSD multiple times. When 2.5 came out EM once again said he believers 2.0 could do FSD and only if it can’t they will upgrade it for free.

Yea in Tesla world EAP is completely done. As long as they have the function it’s done. Despite they call everything beta it’s done. Pretty sure we are living in a parallel world.
 
Just because EM said 10M does not mean it will happen at 10M. Please, we’ve been there.

I don't believe he's the one who said 10 million.

I was simply replying to the poster who mentioned that number of miles (which is you, BTW) as if it's be some long while before they hit it- rather than somewhere between 1-4 weeks.


They’ve said AP2 was capable of FSD multiple times. When 2.5 came out EM once again said he believers 2.0 could do FSD and only if it can’t they will upgrade it for free.

Right, and as before, if it can't, which it can't, the upgrade is free to anyone who bought the feature.

HW3 is where it gets interesting since it's where they've promised specific features will finally be enabled.


Yea in Tesla world EAP is completely done. As long as they have the function it’s done. Despite they call everything beta it’s done. Pretty sure we are living in a parallel world.

Can you cite what features of EAP were in the product description that are still missing once they have drive on nav (without requiring confirmation) and a summon that goes around corners/obstacles on private property?

Because if you can't- then yeah, it's done in every world, as far as adding features.

Doesn't mean they won't improve how well they execute those features over time, as they have continued to do for the earlier EAP features.
 

Waymo is on the verge of "official" deployment of L4 cars (maybe only L3 it's not entirely clear yet till they're "out", in one specific city in Arizona that is very very highly mapped and basically never has rain or any other obscuring weather.

Which is great for that city- but awfully far from a general case.

L3 or L4 on all divided highways, which Tesla seems a lot nearer to than Waymo, would be a vastly preferable feature to L4 in one specific city in AZ.... (unless I suppose you lived in that one city).

I don't expect anybody to get to Level 5 anytime remotely soon.
 
I don't believe he's the one who said 10 million.

I was simply replying to the poster who mentioned that number of miles (which is you, BTW) as if it's be some long while before they hit it- rather than somewhere between 1-4 weeks.




Right, and as before, if it can't, which it can't, the upgrade is free to anyone who bought the feature.

HW3 is where it gets interesting since it's where they've promised specific features will finally be enabled.




Can you cite what features of EAP were in the product description that are still missing once they have drive on nav (without requiring confirmation) and a summon that goes around corners/obstacles on private property?

Because if you can't- then yeah, it's done in every world, as far as adding features.

Doesn't mean they won't improve how well they execute those features over time, as they have continued to do for the earlier EAP features.
Why do I need help you look up the sources and facts?

“Tesla said that owners have already “driven tens of millions of miles to support the validation of Navigate on Autopilot”, but CEO Elon Musk said that they will need about 10 million more miles to make the system automatically initiates the lane changes.”

9E98364B-F8F9-489F-B6DF-775AD4A9DD7E.jpeg


I didn’t say it will be a long while. You are the one who assumed I think it’s going to take long. It will only take days (which is approaching btw). Please re-read what I said.

I said the enhanced part of EAP just started which is also a fact, navigate on AP took 2 years (fact) since they advertised it. It’s JUST out now and is definitely not compete if it still requires confirmation and further validation. Smart summon is not out and is definitely not complete. Don’t forget they also said you could do advanced summon on private roads with AP1 (since removed from site circa 2016, also a fact). How long you think smart summon is out? When it’s out yea we can call it complete, but how long?

I never said they aren’t coming out, but things are happening and won’t be imminent, especially so-called valuable FSD features.

As with all Tesla things I’ll believe it when I see it.
 
  • Informative
Reactions: Tam
Your L software depended on talking to the mothership. If the mothership won’t let the software download, nothing the center can do.

Well, someone on the forum reported they were able to get Tesla to do their L upgrade a few weeks after we had tried to, so I am not entirely sure what the real problem was. But yeah, not much we could do.
 
The general consensus is that FSD will be coming back to order/config at a higher price than before, so is now a good time to double down?

They are just getting started with the E part of EAP with stuff that`s been promised since AP1.
Learn from the past and don´t fall for marketing promises.

Don`t buy nonexistent products from Tesla if you don`t intend to give them an interest free loan anyways.
 
Nothing limits them from removing the option or make it out of stock.

There was a price to add the tow kit to the Model X post delivery. Go to the store and tell me if you can order the part for service center install.

Tesla doesn’t remove upgrades for our benefit.
Hm, OK. But given how heavily they are advertising that every car is ready for full self driving, I cannot really imagine that they would just dump the upgrade option, unless they discover that they aren't technically capable of implementing it (in which case the people who have already purchased FSD wouldn't get it either). The question really is if they would just refuse to honor the price they have given people at the time of ordering so far (i.e. $4k/$5k).
 
I suspect they will honor the $4K/$5K price.....but probably not indefinitely. Tesla might play it nice (like they did just recently with the option of $5K refund OR keep unlimited SC) and when the they reintroduce the FSD option at a higher price of, say, $7,500 to tell existing owners with a previously agreed upgrade option that they have until the end of that quarter to upgrade, otherwise it will be at the new/current post-configuration price of, say $10K.....(prices made up)
 
By the way, I just listened to the Tesla Show podcast for Q3 results. They talk about this subject quite interestingly from minute mark 0:50 minutes for about 15 minutes (to 1:05). They bring up the possibility of Tesla going to a subscription model for FSD.. . Hmmm... Maybe since the hardware is a moving target that can be improved year after year. If they reintroduce FSD as subscription based, I wonder what that would mean for current FSD owners?
 
...current FSD owners?

Owners who paid for FSD fully are covered.

Those who did not pay for it but insist that they got the $4,000 post-delivery price locked in might be out of luck.

A price, even post-delivery price can change at any time because that price was not fully paid.

Tesla verbally said it would honor $4,000 post-delivery price for some grandfathered owners prior to $5,000 increase but again, if owners didn't act on it and paid fully, that price can also change at any time.
 
Owners who paid for FSD fully are covered.

Those who did not pay for it but insist that they got the $4,000 post-delivery price locked in might be out of luck.

A price, even post-delivery price can change at any time because that price was not fully paid.

Tesla verbally said it would honor $4,000 post-delivery price for some grandfathered owners prior to $5,000 increase but again, if owners didn't act on it and paid fully, that price can also change at any time.
That's what I was thinking.
I certainly agree that we have not seen any valuable FSD capability yet, but it sure feels like something is brewing.....
 
...not seen any valuable FSD capability yet, but it sure feels like something is brewing.....

True. It's not FSD yet but the progress is in the right direction.

I used to experience phantom brakes but I haven't with V9. I used to be honked at for lane changing but V9 has improved and it is just so much better. Pre-V9 didn't handle some extra big intersections without lane markers but V9 seems to do very well.

Next is would V9 be able to brake for stationary obstacles while cruising at 90 MPH?

These are still Autopilot levels but FSD is built upon Autopilot's competencies.

By the way, FSD might cost much more than $5,000 just because it might cost much more to work on its software/hardware solutions.

Thus, locking in your FSD price now by paying fully for it might be a good investment against that possibility above.

https://www.eetimes.com/author.asp?section_id=36&doc_id=1333906&page_number=1
 
Last edited:
  • Disagree
Reactions: Phrixotrichus
No. I haven’t seen any plausible FSD features proposed.
People seem to to be counting on some huge breakthrough for some reason. I would look at EAP and judge from that how close the are. It’s not like they have some super secret team who’s way ahead of the EAP team. Now they’re going to start developing software for the new custom computer but I doubt the speed of the neural net is the primary problem they’re facing. I also doubt they’re going to replace all the FSD purchasers computers until they have a compelling reason to.
Elon has said that HW 2.0/2.5 is not capable of running all of the existing cameras at their native (full) resolution and frame rate, and that HW 3.0 will make this possible. Andrej Karpathy has said on earnings calls that the current neural net running in customer vehicles (for EAP) is not indicative or comparable to the neural nets that Tesla has running in the lab. He has said that HW 2.0/2.5 is not capable of running these more advanced neural nets.

We know from Karpathy that HW 3.0 will run a more advanced neural network and I suspect it will provide superior EAP performance and limited FSD “features” from day 1. By this time next year, I suspect much of Tesla’s time/resources will be devoted to developing and improving this more advanced neural network, so that is what I want in my car. An upgrade from the legacy NN to the new, more advanced NN was worth $3000 to me.
 
Last edited: