Starting to regret FSD pre-purchase in a major way

[aaron0k]

You mean this dual motor launch/AP video?

07:25

I guess by your interpretation, he was just saying it could do all those things... just for giggles? Why would he mention those capabilities if there was no intention to leverage them?

I was at the event... in the test drives, the Tesla reps were also telling us that it would be able to stop at signs, lights, exit freeways etc.

You can split hairs on Elon's language, but in the end we've been buying into this Self Driving vaporware for >3 years now...

 

[aaron0k]

Here's an interview with Elon from 2014 where's he's more pragmatic about AP/FSD (@ 02:15).

According to his time-table (2014 + 5-6 years + 2-3 years for reg approval).... you'll have FSD in 2021-2023.

 

[Knightshade]

07:25

I guess by your interpretation, he was just saying it could do all those things... just for giggles?

No, by his actual words he said they could "see" those things.

And they can.

Every thing he says AP can do in that video, it actually does other than the fancy summon bit where it meets you out front because your calendar says what time you're leaving.

Why would he mention those capabilities if there was no intention to leverage them?

AP1 did leverage reading speed limit signs- but it was somewhat hit/miss (and I've heard other brands cars that do this are likely not always reliable at it).

I expect they'll do so again when it's reliable enough.

I was at the event... in the test drives, the Tesla reps were also telling us that it would be able to stop at signs, lights, exit freeways etc.

You can split hairs on Elon's language, but in the end we've been buying into this Self Driving vaporware for >3 years now...

You're mixing up AP and FSD.

Elons talk was not about full self driving. You can tell, because he says autopilot, not full self driving.

And again, AP does everything he promised in that video apart from the fancy summon.

And with drive-on-nav it'll also do exits freeways (which he doesn't mention in that talk), and essentially EAP is feature complete at that point from what was actually promised by Tesla.

FSD is a whole different thing (though obviously leveraging many EAP blocks)

 

[aaron0k]

No, by his actual words he said they could "see" those things. And they can.

Everyone I spoke with at the event left with the impression that AP would obey stop lights/signs and have REAL obstacle avoidance. I'm aware of the difference between AP/FSD. AP1 will never have freeway exists.

I'll concede what was promised, vs what was inferred... however, it WAS inferred.

 

[Knightshade]

Your original claim was

he claimed it would stop at stop signs, exit freeways, etc (watch the Dual Motor/AP event video)

I linked the video. He didn't claim those things at all in that video.

(and exiting freeways wasn't mentioned, AFAIK, until EAP with the AP2 HW- which is going to get it with drive on nav)

I'm not sure you can blame Tesla for people making assumptions about things he didn't actually claim the car would do at the time.

Well- at least not in this case... there's certainly plenty of other cases where Tesla is overly vague about how something is going to work in the future (see for example who gets track mode (and what it'll do) or what's actually different between AWD/P3-/P3+ or a slew of other things).

 

[Bays]

Look around in the industry for non-elon-time predictions. those range from 2025 to 2030+ for level 4/5 and ~2021+ for true level 3.
Tbh I think the only thing that really works well right now is standard TACC.

I don`t expect any early adopters for the Model 3 to ever get anything that`s worthy of the "full self driving" term.

See AP lawsuit history...... it kinda isn`t.


inb4 people clamouring about how their FSD purchase helps Tesla`s "mission"

Most industry is wrong at predicting the speed of disruptive technologies and exponential growth. They keep on revising their predictions forward. Traditional car makers have been very slow at this, the closest is GM because they bought Cruise. However they are still way behind Waymo, which may already be close to being able to drive 100k fully autonomous miles without user intervention.

The problem with Tesla is their set of sensors. The lack of accuracy in identifying lanes, cars, people etc. in v9 software makes it look like they have a long way to go - if the sensor set will ever be able to do unsupervised driving.

 

[aaron0k]

(and exiting freeways wasn't mentioned, AFAIK, until EAP with the AP2 HW- which is going to get it with drive on nav)

June 2014 Shareholders meeting. @ 1:10

"So I think we're making some really good progress on the autopilot side and it, you know I'm confident that in less than a year you'll be able to go from highway on-ramp to highway exit without touching any controls <head nod> (crowd applause)" - Elon​

Was he basing that statement on AP1 or the yet to be announced AP2? Was he leading people to believe AP1 would be capable? Claim or infer?

Not touch any controls? Like gripping your steering wheel every 20 seconds?

 

[csshih]

Was he basing that statement on AP1 or the yet to be announced AP2? Was he leading people to believe AP1 would be capable? Claim or infer?

My guess is AP1. Switching to in-house on AP2 really set them back for quite some time, imo.

 

[babula]

June 2014 Shareholders meeting. @ 1:10

"So I think we're making some really good progress on the autopilot side and it, you know I'm confident that in less than a year you'll be able to go from highway on-ramp to highway exit without touching any controls <head nod> (crowd applause)" - Elon​

Was he basing that statement on AP1 or the yet to be announced AP2? Was he leading people to believe AP1 would be capable? Claim or infer?

Not touch any controls? Like gripping your steering wheel every 20 seconds?

You've mentioned the nag several times in this thread. You do remember that this was specifically added because of drivers being irresponsible with AP right?

If you did a quick Google search you can find several clever workarounds like the water bottle trick, etc...

 

[Knightshade]

Most industry is wrong at predicting the speed of disruptive technologies and exponential growth. They keep on revising their predictions forward. Traditional car makers have been very slow at this, the closest is GM because they bought Cruise. However they are still way behind Waymo, which may already be close to being able to drive 100k fully autonomous miles without user intervention.

The problem with Tesla is their set of sensors. The lack of accuracy in identifying lanes, cars, people etc. in v9 software makes it look like they have a long way to go - if the sensor set will ever be able to do unsupervised driving.

To be fair- Waymo is doing 99% of their miles in one specific place in Arizona where conditions are incredibly simple compared to much of the country.

Being able to drive autonomously in Chandler Arizona isn't nothing, but it also isn't super useful in say Boston or Chicago.

June 2014 Shareholders meeting. @ 1:10

"So I think we're making some really good progress on the autopilot side and it, you know I'm confident that in less than a year you'll be able to go from highway on-ramp to highway exit without touching any controls <head nod> (crowd applause)" - Elon​

Was he basing that statement on AP1 or the yet to be announced AP2? Was he leading people to believe AP1 would be capable? Claim or infer?

Not touch any controls? Like gripping your steering wheel every 20 seconds?

so....completely different presentation than the one you claimed he said it in....(and even then it's said as an off the cuff prediction of future improvements, not what the system will do out of the box- and Drive on Nav essentially fulfills that promise (though obviously the falling out with Mobileye set the timetable back considerably).... Or if you want to be semantic, other than if you want change lanes, you can already go from the end of the on-ramp to the start of the off-ramp on AP without touching any controls (and have been able to for years) beyond weight on the wheel.

The steering wheel thing was already addressed- that was added because people are idiots who can't be trusted.

 

[Bays]

To be fair- Waymo is doing 99% of their miles in one specific place in Arizona where conditions are incredibly simple compared to much of the country.

Sure, it's going to be a long tail. But compare to the incredibly simple conditions that EAP supports.

 

[aaron0k]

so....completely different presentation than the one you claimed he said it in....(and even then it's said as an off the cuff prediction of future improvements,

You mentioned that as far as you knew, he never claimed such a thing. I was just giving you proof that he had, in that same time frame, and well before AP2... semantics?

He said it to Tesla shareholders... (that's not official enough?)... and said he was "confident" (aka "funding secured")... <don't forget the crowd applauding the claim>

Adding the "steering wheel thing", means it wasn't designed well, was rushed, and didn't consider that "stupid people" exist. I also believe Tesla is morally corrupt for not requiring Autopilot safety training before activation.

You can interpret Elon's words anyway you want, but I just see it as a way to justify the constant hyperbole and false claims coming from him. I'll admit, I used to buy into some of it as well...

As it relates to this thread, I've think I've made my case that buying FSD now is either really stupid, or highly philanthropic...

 

[aaron0k]

My guess is AP1. Switching to in-house on AP2 really set them back for quite some time, imo.

...and the ego war between Elon and GeoHot.

 

[COrocket]

I didn’t buy FSD and I see a few issues with the whole situation.

First, let’s say optimistically FSD is 8-10 years away and assuming I still have my car and wanted to purchase it then. If you look up a 10 year old BMW/Audi competitor to compare with average condition and mileage it’s worth $6-8000. Does Tesla really think that the people who waited it out would sink another $6000+ into a car that’s only worth that much? Or just save that money and trade in their old car and put everything towards a newer car purchase.

Second is the premise that the technology will become more valuable to the consumer and therefore Tesla will be able to charge more for it in the future. That’s not how any piece of technology like phones, computers, aircraft avionics etc work. Capability increases and costs go down as a technology reaches the mature stage of its lifecycle. Remember when touch screen nav in a car used to be a multi thousand dollar upgrade offered in high end cars? Now it comes standard with $20,000 econoboxes. I imagine in 10+ years FSD will be offered by many more manufacturers and will have to compete on price, eventually being something that consumers expect to be included in a new car. Tesla probably won’t be able to command $6k+ for their FSD in their older fleet given the pace of technology.

 

[LastGas]

I bought FSD. FSD costs $3,000 ordered today, $5,000 ordered after the car is delivered. The average return on the S&P 500 for the last 20 years is 7.68%. Turning $3,000 into $5,000 at that rate would take you 7 years. Do you really think FSD will take 7 more years? Are you confident that the price won't increase again?

Another comment suggested that FSD wouldn't be available before the car falls apart. How long do you expect your Model 3 to last? Surely it's good for 200,000 miles if not a lot more. At 15,000 miles per year, that comes to 13 years. Do people think FSD will take 13 years? I plan to keep my Model 3 at least that long.

Finally, even if FSD is 2-3 more years, bits and pieces of it will be out before then. Isn't one of the perks of owning a Tesla being on the bleeding edge of new technology? I think so.

 

[KenF]

In early August, Elon said the Autopilot 3.0 hardware would debut with the first FSD-exclusive software features in 4 to 6 months (i.e., 6 to 12 months in Elon speak), and would be a free upgrade for those that purchased the FSD option.

I paid for FSD because I love technology and I want access to all new Tesla software features / improvements. You don’t get that without the FSD option.

 

[Daniel in SD]

Buy options in Tesla stock. Buy $3k Jan 15th 2021 options @450 for $30 each. If Tesla actually gets FSD working the stock will be $1000. Then you’ll have $55k to buy a brand new Tesla (after trade in of course).
A perfect plan

 

[aaron0k]

In early August, Elon said the Autopilot 3.0 hardware ... would be a free upgrade for those that purchased the FSD option.

I commend Tesla doing the right thing... however, from another perspective; one could consider that a 'recall'. Nothing is free.

 

[LM-5]

I commend Tesla doing the right thing... however, from another perspective; one could consider that a 'recall'. Nothing is free.

How did you parlay "brand new hardware with increased power and capabilities" into a "recall"?

At any rate, this entire thread is "I'm going to justify my purchasing decision by telling people how ignorant they are for not making the same decision I made." Buy it, don't buy it, whatever. To each their own.

 

[daniel]

I waffled quite a bit but ultimately decided to plunk down the pre-order cash for FSD functionality “when it becomes available.” I knew this was a huge risk, and I was told by several industry insider friends that I shouldn’t, that Tesla - or anyone else - is still many years away from truly delivering this in the way someone would expect when they say “full self driving autonomy.” I did it anyway.

I’ve been patiently waiting for v9, knowing that the “drive on nav,” while clearly not anything close to FSD, would at least be a big step in the right direction.

Now, after over 2 months of teasing a v9 release date only to delay it, tease it again, then delay it some more, Musk comes out a few hours ago to say they’re pulling drive on nav “for a few more weeks.” Not only that, but in doing so, he also mentions that it’s “extremely difficult to achieve a general solution that works well everywhere.”

This is a serious admission of just how far off Tesla is from FSD. It’s that difficult for them to get freeway “semi-autonomy” (until this car stops nagging me every 20-30 sec to grab the wheel, nothing they do is autonomous IMO) working, and freeway driving is FAR less complicated than any other roadway.

I am seriously questioning at this point whether FSD will be available at any point before my car falls apart due to age alone. If that were to be the case, how can it be legal for Tesla to charge customers money up front for a feature they never deliver before the product deprecates beyond usefulness?

I did not pay for FSD because I know that full self-driving is at least a decade away and by then I'd want a newer car anyway. But then I realized that the "FSD package" is not synonymous with full self-driving. Presumably Tesla will start including some features in FSD package cars that EAP package cars do not get. I'd consider recognition of stop signs and stoplights to be worth the upgrade. But when those features will come, I have no guess.

I bought FSD. FSD costs $3,000 ordered today, $5,000 ordered after the car is delivered. The average return on the S&P 500 for the last 20 years is 7.68%. Turning $3,000 into $5,000 at that rate would take you 7 years. Do you really think FSD will take 7 more years?

Yes.

I think that full self-driving will take 10 to 15 years. My guess is Level 4 autonomy in ten years and Level 5 in twenty years. These are very difficult things to achieve.

However, as noted above, maybe they can get to Level 3 autonomy in 5 years, and that would be worth the price of the FSD package. Even just stop sign and stoplight recognition might be worth it to me, as noted above.

Automatically taking freeway off ramps based on NAV is not worth much to me, but making turns on city streets based on NAV would be, though I think that's a long way away.

The other reason I didn't pay for FSD is that I think that an ideal implementation will require more sensing hardware than our cars have. I'd rather buy a new car in ten years, with the ideal hardware, than accept an implementation that meets minimum requirements using the present sensors.

Note that my car now (EAP and v9 firmware) is the most sophisticated autopilot system available to the consumer today. FSD package cars may get features I don't in the future, but I'm skeptical that those will be worth $3,000 to me any time in the next 5 to 7 years.