$12K for FSD is insane

[raptor5244]

How people keep falsely claiming they're selling something that doesn't exist is mind boggling to me.

Especially when this factually untrue claim has been explicitly debunked repeatedly in this very thread.

6 of the 7 promised features of the current FSD package are in wide release, and have been for years.

The 7th exists, but is still in narrow-release beta testing. No shortage of videos out there showing owners testing it, so that it exists is self evident.

I think it's about setting proper expectations and I don't think Tesla has done a good job of doing so when it comes to this software. I realize it is a work in progress but many people expect the car to "full self drive" or at least be trustworthy and usable enough where buyers feel they are getting some value out of the package when spending this kind of money. Technically, a feature like Auto Park may be delivered but as you mentioned yourself it is slow and unreliable for perpendicular parking. So, if that is the case then you will typically not use the feature. If you can't trust the system to park how can you trust it to full self drive on a busy street?

For me, FSD and all of the associated features need to work as good or better than I can do it myself or it just winds up being annoying and frustrating to use. If or when AI reliably works as good or better than I can do it myself then I will pay but nowhere near $12k since I enjoy driving. Until then I would happily beta test for free or even pay a $1000 and I would treat it as a hobby. But it would be just that, a hobby since I don't really care if the car can full self drive or not.

 

[ryanbush]

I agree, it's not worth it. Maybe Musk is trying to recoup future hardware costs.

But I have to admit it's surprising how ridiculously fast those end of quarter FSD equipped (reddit inventory) vehicles sold given all the issues.

If you purchase a vehicle in inventory that shows FSD, they can remove it during the purchase process

 

[Knightshade]

But they are not. This just is not true

Everything you quoted is absolutely true.

6 of the 7 listed features are already wide release/delivered, and have been for years.

The 7th is in limited-release beta testing.

Which part is "not true"?

One, you have to first EARN the right to be put in the beta by getting a 99+ score. THEN IF you live in an area they decide to pick for you then you can use a system that is less than 100% delivered as promised.

Yes- that's the limited release beta.

Being in the beta and testing the pre-production version is not a promised element of buying FSD

Receiving the wide-release delivered product is. That isn't finished yet.

Let's be VERY clear it IS NOT in wide release and hasn't been for years.

You appear to have misread my post and are arguing about something I never wrote.

Which is weird since you quoted the text while still misreading it.



Here it is again.

These are facts.



6 of the 7 promised features of the current FSD package are in wide release, and have been for years.

The 7th exists, but is still in narrow-release beta testing. No shortage of videos out there showing owners testing it, so that it exists is self evident.

 

[Knightshade]

I think it's about setting proper expectations and I don't think Tesla has done a good job of doing so when it comes to this software. I realize it is a work in progress but many people expect the car to "full self drive"

So for pre-March 2019 buyers the description was very broad-- and I agree those folks have a right to expect a car that will, in most cases anyway, actually "self drive"

But that's not the version we're discussing here where the price is going up to 12k.

The post 3/19 version, nearing 3 years old now, is VERY clear about what features it includes, and at what level of automation.


If folks don't bother to read a bullet point list of what they're paying for, that's really on them.

And thanks to the subscription model if they're unsure, they can invest $199 to try it out for a month too if they DID read it and are still unsure.

 

[Supcom]

How Tesla can continue to legally sell something that doesn't exist is mind boggling to me...

Ever heard of kickstarter?

 

[goin2drt]

@Knightshade I guess we will agree to disagree. It is called Full Self Driving. You do not get FSD for your now $12K. It is in very NARROW release to beta testers that get hand picked based on a score and location that they live. For $12K it is a waste of money until it is actually a fully functional and EVERYONE (wide availability as I think you stated) can have it.

You obviously love it and believe in the value. I do not so therefore I will not be buying. Cheers and enjoy your FSD that is what makes it a great option. Those that want it can buy it. Enjoy.

 

[Sporty]

…. 6 of the 7 promised features of the current FSD package are in wide release, and have been for years. …

But some of them “don’t work”. They are marked as ‘beta’ so it’s ok that they have some problems/bugs/usability-issues. Did they really deliver on those features yet?

 

[Sporty]

How Tesla can continue to legally sell something that doesn't exist is mind boggling to me... My guess here is because they moved the actual working features under FSD (auto-park/autopilot being the main things) they can claim it's not actually vaporware... All I know is Elon is a liar, and don't give me this "he just overestimates the capabilities of his team" crap, he lies. He's been lying since the first Model S rolled off the line and the list of lies now is quite long. I just wonder how much longer he can keep this charade up before people figure it out and Tesla "evaporates" almost overnight. Then again, I've wondered that for the better part of 4 years and Tesla continues to grow by leaps and bounds so who knows. Clearly not me... LOL

Jeff

They’re not going to evaporate. They have a lot going on that is good. They may end up course correcting and needing to do something for some number of customers. I doubt it would be a full refund… people got some use of something for years etc. (I’d take a coupon .

 

[chinney]

Years of actual data of publicly released features show this claim is fundamentally untrue.

AP has been this way since launch (an L2 system requiring an attentive driver), and we have year after year after year of data showing lower accident rates using it than not.

By removing the tasks of making all the little micro adjustments to steering and throttle for long highway trips for example your brain has morebandwidth to actually pay attention to what's going on around the vehicle.

Further, since the system can see 360 degrees at all times, and you can't, it'll catch things you wouldn't.




You again seem to be confusing a limited-release beta test with the actual public FSD offering that has been in wide release for years now and works very well and is the thing whose price is increasing to 12k per the thread title.

Tesla Vehicle Safety Report | Tesla

At Tesla, we believe that technology can help improve safety. That’s why Tesla vehicles are engineered to be the safest cars in the world. We believe the unique combination of passive safety, active safety, and automated driver assistance is crucial for keeping not just Tesla drivers and...

www.tesla.com

While there's a valid argument to be made comparing AP miles to non-AP miles is not precisely apples to apples, the difference in accident rate is still far lower for a Tesla on AP than other vehicles.


So again your premise a good L2 system makes the car less safe is directly contrary to all available evidence.

And that information comes from...um...Tesla. There are reasons to doubt the accuracy and relevance of that data for reasons that are dealt with, in part and for example here: Elon Musk says a Tesla on Autopilot is 10 times less likely to crash than the average car, but experts say that stat is misleading. Perhaps for people who are, in effect, committed beta testers prepared and mentally able to dedicate themselves to the special type of monitoring required, it might well be safer. But for the average driver, I doubt it. I have no objection to Tesla continuing with this long-term beta test - which in the long-term, with further hardware, software and connectivity improvements, probably will bear fruit - but let's call it what it is.

 

[Knightshade]

And that information comes from...um...Tesla. There are reasons to doubt the accuracy and relevance of that data for reasons that are dealt with, in part and, for example, here:

Literally nothing in your link provides any reason to doubt the accuracy of the data.

They're pointing out the same thing I already did- that it's not apples to apples because AP tends to be used more often on highways.

But even adjusting for that the crash rate is still lower than average when using AP....(for example the story notes roughly a 5:1 accident ratio of local vs city fatal accidents.... yet the Tesla does 10x better overall).

Bonus points for a source that cites Missy Cummings, an 'expert' whose bias is so well known she was forced to recuse herself from any NHTSA work related to Tesla... (she was on the board of directors of a LIDAR company so has published a lot of anti-tesla FUD in the past)


Seems a bit like your claim it's "more dangerous", without having any evidence to support such a claim.

Perhaps for people who are, in effect, committed beta testers

You seem to again be confusing the publicly released AP (what the tesla data, and your link are discussing) and the FSDBeta program which is not at all relevant to that.

But for the average driver, I doubt it.

Tell me you've never used it without telling me you've never used it

Where's your data again?

@Knightshade I guess we will agree to disagree. It is called Full Self Driving.

McDonalds sells a thing called a Happy Meal, but I don't ask for a refund if the kid eating it is still in a bad mood afterward.

A Diaper Genie doesn't grant wishes.

A radio flyer neither receives radio NOR flies. OMG THE FRAUD!

Also spoiler- Plaid Mode in the Model S doesn't change the paint color.



Tesla literally has a link spelling out what the FSD package does during the purchase process-- if you buy it and are shocked to find out only later what's in it- that's on you.

But some of them “don’t work”.

All of them work (the delivered ones I cite)

If they work well enough to be worth the price of buying them is up to the individual buyer of course.

One example I know I mentioned- basic summon has essentially 0 value to me, but others have posted they use it all the time (often folks with say a really narrow garage)

Likewise the more advanced summon has been useful to me to have my car pull under a covered area near the door at my workplace when it's raining--- others haven't found it useful at all because of its limited range.

For someone who does nearly all their driving in a city and on local dense roads, NoA isn't useful. For someone doing primarily highway driving it's phenomenal

Hence again why it's great anybody can find out for themselves for just $199 up front for a month- testing the various features in their own specific conditions, locations, and circumstances.

 

[nvx1977]

There's another detail here. People who talk about enjoying "driving" are living in the past. Way in the past.

Several people have hit it on the head. Driving from point A to Point B is now more like individual cabins hitching up to a train. The freeway. The highway, the city Boulevards, and increasingly City Streets, that are widened as much as possible. There will soon be no excuse for coupling sloppily to the caterpillar - herky jerky, bumper to bumper. And if you want to "have fun" outside the circus track, you will have to be "original", reckless, unpredictable, cause accidents and collect tickets. Judging by the crazy accidents and fatalities among human operators, it's obvious how lousy people are are cooperatively riding that train.

I'm sorry, I realize not everybody is loving the present. There are still a few hillsides to chew up with a 4x4, and some suburban spots that aren't covered in parking lots full of plastic franchise shops, connected by ribbons of asphalt. It's not all that lovable. But that's the trend.

The whole idea of "freedom" at the controls of a free-wheeling internal combustion machine is over. Like a Marlboro commercial. A Norman Rockwell painting. Like flying in barnstorming days. Now, flying is all about ALS, Glide-slope, from TCA to TCA, using autopilots of different types. Funny how things change.

I don't care. Can't beat 'em, join 'em. Look at those fat ribbons of tail-lights from above! Is that what y'all consider "fun driving"? Electric propulsion is a better match, and pollutes less. Automatic steering, lane-switching, merges, entries, exits. We're all getting older too. Don't deny it. It's rather relaxing to hand over part of the job.

I swear, the people who are now bitching about how the Tesla system won't let them sleep in the back will look back longingly at the days when there was still something fun and imperfect about operating the machine. I'm still enjoying it.

It's pretty presumptuous of you to think people who enjoy driving aren't future-looking or adaptable to change.

I like carbonara. It's a centuries-old Roman-Italian dish. I must be living in the past.

Almost everyone I've talked to who's gotten a Model 3 has said it's really fun to drive. Small steering wheel, smaller steering ratio, low center of gravity, RWD, etc. If you don't get it, that's fine. No need to criticize others.

 

[Knightshade]

BTW another data point on the nonsense suggesting AP is more dangerous, rather than accepting the data is safer than just human driving.

AP keeps improving. So even if you ignore the apples to oranges issues, the AP accident rate keeps declining over time as the system gets better even comparing JUST cars running on AP.

https://twitter.com/x/status/1483633820813385730

 

[raptor5244]

BTW another data point on the nonsense suggesting AP is more dangerous, rather than accepting the data is safer than just human driving.

AP keeps improving. So even if you ignore the apples to oranges issues, the AP accident rate keeps declining over time as the system gets better even comparing JUST cars running on AP.

https://twitter.com/x/status/1483633820813385730

What specifically about AP are they attributing this to? Staying in your lane? Maintaining a constant speed and distance? I am not being coy, I am genuinely curious as to how it reduces accidents. AP is only supported on the highway, so I assume they are referring to highway based accidents only? People dozing off at the wheel? Drunk driving?

 

[raptor5244]

It's pretty presumptuous of you to think people who enjoy driving aren't future-looking or adaptable to change.

I like carbonara. It's a centuries-old Roman-Italian dish. I must be living in the past.

Almost everyone I've talked to who's gotten a Model 3 has said it's really fun to drive. Small steering wheel, smaller steering ratio, low center of gravity, RWD, etc. If you don't get it, that's fine. No need to criticize others.

Now I am hungry for some carbonara. I had an M3P and SR+. While the M3P is obviously the faster car I actually find the SR+ more fun to drive since it is lighter and RWD only. With the rear motor only it has less dive into the corners and changes direction noticeably quicker. I hope Tesla comes out with a RWD only performance trim with sporty seats with proper bolstering, mag ride suspension, HUD, Track mode, flare the fenders a bit more and put a more aggressive wheel and tire package on. Price it at $50-55k.

 

[OxBrew]

What specifically about AP are they attributing this to? Staying in your lane? Maintaining a constant speed and distance? I am not being coy, I am genuinely curious as to how it reduces accidents. AP is only supported on the highway, so I assume they are referring to highway based accidents only? People dozing off at the wheel? Drunk driving?

AP is so poor at dealing with traffic at speed in corners (at least vision only) that people are increasingly only using it in really straight, open predictable stretches.

Heck, in the 300 mile trips across the state, I only use it for maybe 75 miles. So the AP gets the benefit of the safer parts of the highway, I take all the risk in the tougher parts.

The total percentage of miles driven on AP for the fleet could be 1/3 what it was 2 years ago, and accidents per mile could easily look better, like they apparently do.

No one outside Tesal can confirm or deny this. They are extremely careful to never qualify their "accidents per million miles" data with any context that could clear any of this up.

If they are willfully obscuring this context, and cherry picking the data, it would be clearly unethical and unwise, if the truth is that AP can't actually handle dangerous situations (it can't, at all, in my 6000 miles of experience).

If they don't actually understand how the end user is naturally taking over in dangerous situations, leaving AP to do the safe easy driving, they are not qualified to comment on AP safety.

From out here, we can't know the actual truth, but I know what my guess is.

 

[Raurele]

AP is so poor at dealing with traffic at speed in corners (at least vision only) that people are increasingly only using it in really straight, open predictable stretches.

Heck, in the 300 mile trips across the state, I only use it for maybe 75 miles. So the AP gets the benefit of the safer parts of the highway, I take all the risk in the tougher parts.

The total percentage of miles driven on AP for the fleet could be 1/3 what it was 2 years ago, and accidents per mile could easily look better, like they apparently do.

No one outside Tesal can confirm or deny this. They are extremely careful to never qualify their "accidents per million miles" data with any context that could clear any of this up.

If they are willfully obscuring this context, and cherry picking the data, it would be clearly unethical and unwise, if the truth is that AP can't actually handle dangerous situations (it can't, at all, in my 6000 miles of experience).

If they don't actually understand how the end user is naturally taking over in dangerous situations, leaving AP to do the safe easy driving, they are not qualified to comment on AP safety.

From out here, we can't know the actual truth, but I know what my guess is.

I use autopilot 95% of the time. drove 350 miles today and the only times I didn't use it were when I had to change over multiple lanes quickly to get to a turn or exit. give me a break. you must be doing something wrong.

 

[raptor5244]

AP is so poor at dealing with traffic at speed in corners (at least vision only) that people are increasingly only using it in really straight, open predictable stretches.

Heck, in the 300 mile trips across the state, I only use it for maybe 75 miles. So the AP gets the benefit of the safer parts of the highway, I take all the risk in the tougher parts.

The total percentage of miles driven on AP for the fleet could be 1/3 what it was 2 years ago, and accidents per mile could easily look better, like they apparently do.

No one outside Tesal can confirm or deny this. They are extremely careful to never qualify their "accidents per million miles" data with any context that could clear any of this up.

If they are willfully obscuring this context, and cherry picking the data, it would be clearly unethical and unwise, if the truth is that AP can't actually handle dangerous situations (it can't, at all, in my 6000 miles of experience).

If they don't actually understand how the end user is naturally taking over in dangerous situations, leaving AP to do the safe easy driving, they are not qualified to comment on AP safety.

From out here, we can't know the actual truth, but I know what my guess is.

Good point regarding AP doing the easy stuff while you take on the the more complicated situations. I think there is some "Hey look, no hands" during the wide open stretches but when you are dealing with heavier traffic that may need a quick reaction we tend to lean more on our experience and anticipate what is going to happen rather than waiting to see if AP is smart enough to handle it yet. That kind of trust will take a while.

 

[Knightshade]

What specifically about AP are they attributing this to? Staying in your lane? Maintaining a constant speed and distance? I am not being coy, I am genuinely curious as to how it reduces accidents. AP is only supported on the highway, so I assume they are referring to highway based accidents only? People dozing off at the wheel? Drunk driving?

The stat is all uses of AP. Officially that's intended to be highway, but there's myriad threads of people still using it elsewhere (and then often complaining when it does something that don't like because they used it somewhere they shouldn't have)-- still that data would be included in the data set.

As to why AP is safer I think you already gave some of those answers... if you're sleepy or tipsy or tired or whatever it's obviously gonna do a better job maintaining lane and speed/distance than the human will alone.... And for some drivers it'll be better at those things even fully alert (minimum follow distance is still not as close as I've seen some humans tailgate).... see also the idiots on their phones who smash into the rear of a car in front on the highway they didn't notice had slowed down.

Additionally, the system sees all directions at once, and without the side mirror blindspot a human may have- so it's going to catch at least SOME unsafe lane changes a human might not.

I've also personally had it move itself over quicker than I could react when a truck started to drift into my lane.


None of this is to say it'll always do all those things 100% flawlessly...nothing is perfect. But any time it does any of them that's additional safety a "dumb" car does not benefit from.



That said I'm a little confused by the "heavy traffic" remarks... assuming you mean on highways AP is phenomenal at this task. Far better than any active cruise/lane keep system I've used from Toyota, Hyundai, Ford, GM, Honda, Lexus, etc...

 

[Sporty]

I use autopilot 95% of the time. drove 350 miles today and the only times I didn't use it were when I had to change over multiple lanes quickly to get to a turn or exit. give me a break. you must be doing something wrong.

I agree with OxBrew. The stats are are suspect. I too barely use AP. Only in certain situations which are the safe and simpler parts of a drive. If many others are doing that, then the stats are skews. And I agree only Tesla a knows if that’s the case or not.

Why I only use AP is some case is immaterial to the fact that I do.

 

[OxBrew]

That said I'm a little confused by the "heavy traffic" remarks... assuming you mean on highways AP is phenomenal at this task

I mean crowded highway, but still moving at speed limit and higher, with trucks and curves (like I90 in Washington going over Snoqualmie Pass). AP is dangerous, not fit for purpose.

Vision Only.