FSD Timeline Promises (summary)

[Knightshade]

The problem with this is it gave the impression that it was ready for validation.

Their premise then was "We have a system that will work, as soon as we feed it enough fleet data"

This model failed validation- adding fleet data was insufficient.

They've re-worked the system several times since then- each time believing the new approach would solve the issues the previous one ran into.

Validation isn't just "Ok, this is TOTALLY DONE ALREADY and we just need to make sure it's stable"

It's "Does this thing we think will be able to do X, Y, and Z actually do X, Y, and Z.

If the SW is 'ready" there's nothing else to validate. Validation comes before the SW is ready.

It's pretty common to find problems in this process (it's WHY you do it)- and then have to make changes to the SW to address them.

In this case the changes were extensive- and happened multiple times.

They're still doing that as they find new issues- but believe the current overall approach makes it solvable.

Are they right this time? I don't know. Neither do you.

Just because a person doesn't buy FSD doesn't mean that FSD wasn't part of the decision to buy the vehicle.

It did say it was capable of full self driving which at the very least hints at more advanced driving capabilities than anything else available for purchase.

And that's still true today.

Other OEMs systems are generally functionally inferior, and more restricted in where they can even be used. Even versus the wide-release version of FSD today.

Not relevant to Tesla fans who disagree with the way FSD is marketed.

Since 3/19, when they spelled out the exact features you'd get and spelled out it was being sold as L2?

Or are you now moving the goalposts to "We are mad about how they USED to market it years ago and... want to yell at the clouds some more about it..."?

I do not believe there are people who believed Tesla would achieve FSD and did not pay $3k for it.

Then your belief is factually wrong. See below for at least 2 examples.

Back in 2018 when I bought FSD for $3K I felt like most people buying it did so roughly for the same reasons I did. They didn't actually think Tesla would achieve FSD, but had other reasons.

Bingo.

I still don't think they can actually get the L5 with current HW (though with sufficient NN improvement I don't think they'd need a lot more HW, probably just 2 or maybe 4 added cameras low and side/front facing on the fenders- but a higher-compute driving computer).

But I bought for 3k because I believed they absolutely COULD get to at least L3 highway- and probably L4 highway. And that it'd probably require more compute which I was promised I'd get for free if I bought. So I did.

If I get L5 eventually-cool. But if I "only" get L3 or L4 highway I'm 1000% happy with my purchase.

 

[daniel]

Then you have two choices. Wait for it to be delivered, or sell the car and take Tesla to court/arbitration over retaining the full amount for FSD without fully delivering during the time of your ownership.

Twister: If you had a plan drawn up and formed the business to rent out your robotaxi, you might not have to wait to dispose of the car. But "expectation damages" may be a remedy in that specific case. My fellow attorney active on the boards may have additional wisdom to dispense.

On paper, filing a lawsuit is a possible remedy. But our legal system being what it is, you have a near-zero chance of winning against a large corporation with a whole team of high-paid lawyers using their entire work time preparing for just such lawsuits. And as noted, I have no standing to sue since I didn't buy FSD.

I have to admit to being pretty baffled by the line of reasoning that goes I AM REALLY MAD ABOUT NON DELIVERY OF THE THING I NEVER BOUGHT.

Seems to make some people happy though.

I am upset that my company (I own shares because they make the best cars in the world) has stooped to such deceptive practices. I am not going to divest myself over this, because they still make the best cars in the world. But I'd like them to return (with interest) the money they obtained under false pretenses.

I do not believe there are people who believed Tesla would achieve FSD and did not pay $3k for it.

I fully believed that Tesla would achieve true FSD. I did not believe that they would achieve it during the time I expected to own my car. And I did not believe that when true FSD finally existed, that it would be possible to retrofit my car with the needed hardware. Tesla claimed the cars had all the hardware that would be needed to run software that did not yet exist! You cannot know what hardware will be needed until you have the software.

When and if Tesla has true FSD for purchase by the public, then I will buy it. Since I expect that to be 5 to 15 years from now, I'll trade in my car for the FSD model. If I'm still alive. At my age I could die at any time.

And if another company achieves it first I'll probably buy that car. I hope it's Tesla.

 

[Daniel in SD]

Since 3/19, when they spelled out the exact features you'd get and spelled out it was being sold as L2?

Or are you now moving the goalposts to "We are mad about how they USED to market it years ago and... want to yell at the clouds some more about it..."?

I don't think I'm mad about it. The new disclaimers are better but I don't think they're clear enough for many people. It doesn't help that Tesla sells a product called FSD that may always require a driver ("dependent on achieving"), sold driverless FSD in the past, and has a future robotaxi product also called FSD.

Then your belief is factually wrong. See below for at least 2 examples.

I said that I don't believe this person exists:

• believed Tesla would deliver driverless FSD to their vehicle
• did not think it was worth $3k

 

[AlanSubie4Life]

It doesn't help that Tesla sells a product called FSD that may always require a driver ("dependent on achieving"), sold driverless FSD in the past, and has a future robotaxi product also called FSD.

I feel this is an excellent succinct summary of the types of FSD. Does anyone disagree with this characterization, I wonder?

 

[Daniel in SD]

I feel this is an excellent succinct summary of the types of FSD. Does anyone disagree with this characterization, I wonder?

I think the real controversy is over what FSD Beta is. I think it's a beta of FSD robotaxi. Some people think it's a beta of "autosteer on city streets (beta)." It's not clear where the briefly sold "automatic driving on city streets" FSD feature figures into all this.

 

[Knightshade]

I think the real controversy is over what FSD Beta is. I think it's a beta of FSD robotaxi. Some people think it's a beta of "autosteer on city streets (beta)." It's not clear where the briefly sold "automatic driving on city streets" FSD feature figures into all this.

Read the CA DMV docs. It's very very clear.

What is currently being tested is the L2 city streets code.

It's the last thing owed to post March 2019 FSD buyers.


If you want me on board anyplace with "confusing" naming that's my stop.

It should've been called City Streets Beta. Because that's what it is.

 

[rxlawdude]

On paper, filing a lawsuit is a possible remedy. But our legal system being what it is, you have a near-zero chance of winning against a large corporation with a whole team of high-paid lawyers using their entire work time preparing for just such lawsuits. And as noted, I have no standing to sue since I didn't buy FSD.



I am upset that my company (I own shares because they make the best cars in the world) has stooped to such deceptive practices. I am not going to divest myself over this, because they still make the best cars in the world. But I'd like them to return (with interest) the money they obtained under false pretenses.



I fully believed that Tesla would achieve true FSD. I did not believe that they would achieve it during the time I expected to own my car. And I did not believe that when true FSD finally existed, that it would be possible to retrofit my car with the needed hardware. Tesla claimed the cars had all the hardware that would be needed to run software that did not yet exist! You cannot know what hardware will be needed until you have the software.

When and if Tesla has true FSD for purchase by the public, then I will buy it. Since I expect that to be 5 to 15 years from now, I'll trade in my car for the FSD model. If I'm still alive. At my age I could die at any time.

And if another company achieves it first I'll probably buy that car. I hope it's Tesla.

So you're cool with share price depressing news of Tesla "admitting" to violating the alleged law?

Remember Tesla's corporate duty runs to its shareholders.

 

[daniel]

So you're cool with share price depressing news of Tesla "admitting" to violating the alleged law?

Remember Tesla's corporate duty runs to its shareholders.

I don't think the FSD kerfuffle has a significant effect on share price. I suspect that the real market movers (large funds) don't expect Tesla to achieve real FSD any time soon. I think that the large swings in share price that have characterized TSLA since its inception have other causes entirely and would dwarf any effect of Tesla saying, "We made a mistake in promising FSD before we had a good handle on the difficulties involved. For that reason, we are offering a full refund plus five percent annual interest to any people who bought FSD before <date> [the date when Tesla stopped promising what effectively amounted to L5]."

I am a shareholder. I want "my" company to be ethical. Tesla does not need to misrepresent its progress toward a driverless car to be successful and see a long-term rise in share value. As a shareholder, I feel that Tesla's duty to me is to run a completely honest business and to continue making the best and safest cars in the world, as well as equipment for solar power.

 

[t3sl4drvr]

Just because a person doesn't buy FSD doesn't mean that FSD wasn't part of the decision to buy the vehicle.

I did buy my ~$80k Tesla Model S based on the advertised fact that it is capable of driving itself after "extensive validation" and that is includes all hardware needed for full self driving. I wanted to be be one of the first to to own self-driving car. While it was more expensive, I postponed FSD software purchase until Tesla gave a credible timeline for delivery. (end of 2020) and ensured I had FSD computer installed ASAP to be one of the first to get the Navigate on City Streets feature.

 

[t3sl4drvr]

So you're cool with share price depressing news of Tesla "admitting" to violating the alleged law?

Do you hold a position in TSLA?

Are you spreading misinformation here in order to help preserving share value?

 

[S4WRXTTCS]

Read the CA DMV docs. It's very very clear.

What is currently being tested is the L2 city streets code.

It's the last thing owed to post March 2019 FSD buyers.


If you want me on board anyplace with "confusing" naming that's my stop.

It should've been called City Streets Beta. Because that's what it is.

What's clear is what's being tested, and not what's intended.

Regardless of what people feel they promised we all can agree on what's intended, and what's intended certainly isn't L2.

What's being tested is actually a combination of things. Right now its largely L2 city streets, but for most of us its also our first exposure to Tesla Vision without the Radar. Where it's noticeably more beta than what we had before.

It's so bad that I have zero concern about moving to a single stack, and once we move to a single stack then its really FSD Beta.

 

[EVNow]

I do not believe there are people who believed Tesla would achieve FSD and did not pay $3k for it.

I said that I don't believe this person exists:

• believed Tesla would deliver driverless FSD to their vehicle
• did not think it was worth $3k

Plenty - people who don't grasp the implications of FSD.

My neighbor for one, when I talked to him in 2019 (around the time I bought FSD for 2k !).

 

[Daniel in SD]

Plenty - people who don't grasp the implications of FSD.

My neighbor for one, when I talked to him in 2019 (around the time I bought FSD for 2k !).

He thought the car would be able to drive him while he was in the backseat and he didn't think it was worth paying $3k for that?

 

[EVNow]

He thought the car would be able to drive him while he was in the backseat and he didn't think it was worth paying $3k for that?

Right.

In his case - as a s/w guy - he thought the system would always have bugs and didn't want to take any chances. I brought up the fact that the system could be more reliable than a human driver at some point - his thinking was, if he made a mistake he was prepared for it - but didn't want an accident because the "s/w" made a mistake.

In any case, all I'm saying is - people could have all kinds of reasons for not getting FSD - even if they believed the car would get it. Infact in 2018/19 when I talked about FSD to a number of my engineering colleagues - most of them had "no doubt" Tesla can achieve FSD in a year or two. But few showed any interest in actually getting it (some had Tesla, some didn't but were planning to get the car - but not FSD).

The estimates say about 10 to 15% get FSD now - we can't assume rest 85 to 90% don't think FSD will be achieved soon enough.

Funnily a lot of us who bought FSD, don't think FSD will be achieved soon !!

 

[Daniel in SD]

Right.

In his case - as a s/w guy - he thought the system would always have bugs and didn't want to take any chances. I brought up the fact that the system could be more reliable than a human driver at some point - his thinking was, if he made a mistake he was prepared for it - but didn't want an accident because the "s/w" made a mistake.

In any case, all I'm saying is - people could have all kinds of reasons for not getting FSD - even if they believed the car would get it. Infact in 2018/19 when I talked about FSD to a number of my engineering colleagues - most of them had "no doubt" Tesla can achieve FSD in a year or two. But few showed any interest in actually getting it (some had Tesla, some didn't but were planning to get the car - but not FSD).

The estimates say about 10 to 15% get FSD now - we can't assume rest 85 to 90% don't think FSD will be achieved soon enough.

Funnily a lot of us who bought FSD, don't think FSD will be achieved soon !!

But with FSD you don't even have to be in the car! You would never have to look for parking. You could have the car go pick up something from the store. I must admit I'm mystified.

 

[EVNow]

Their premise then was "We have a system that will work, as soon as we feed it enough fleet data"

This model failed validation- adding fleet data was insufficient.

They've re-worked the system several times since then- each time believing the new approach would solve the issues the previous one ran into.

This is a good explanation.

I think people who are angry with Tesla/Elon assume Elon was lying when he said FSD would be ready by end of the year etc. They don't seem to be able to accept (what I think is) the fact that Elon was simply wrong. Multiple times.

Infact a lot of people think he has always been completely wrong, even now and only Lidar and HD maps can solve FSD.

Ofcourse, there is a significant overlap between the above two groups (people who think world's richest man is a fraud/charlatan).

 

[EVNow]

But with FSD you don't even have to be in the car! You would never have to look for parking. You could have the car go pick up something from the store. I must admit I'm mystified.

People are not rational - they only rationalize their decisions made on "gut feeling".

How else would you explain a rural drug addict, who talks and texts when driving - supporting wars in far away places because those brown people might come and harm her in some way ? People have very poor concept of risk & rewards.

 

[S4WRXTTCS]

Right.

In his case - as a s/w guy - he thought the system would always have bugs and didn't want to take any chances. I brought up the fact that the system could be more reliable than a human driver at some point - his thinking was, if he made a mistake he was prepared for it - but didn't want an accident because the "s/w" made a mistake.

In any case, all I'm saying is - people could have all kinds of reasons for not getting FSD - even if they believed the car would get it. Infact in 2018/19 when I talked about FSD to a number of my engineering colleagues - most of them had "no doubt" Tesla can achieve FSD in a year or two. But few showed any interest in actually getting it (some had Tesla, some didn't but were planning to get the car - but not FSD).

The estimates say about 10 to 15% get FSD now - we can't assume rest 85 to 90% don't think FSD will be achieved soon enough.

Funnily a lot of us who bought FSD, don't think FSD will be achieved soon !!

I think its highly dependent on the type of Engineer, and their experience..

In my 25+ years as an Engineer I've only had a couple of occasions where I had to design a system that didn't allow for a single fault failure. Where I had to include failsafe HW.

The lack of redundancy at the HW level was really at the heart of my own disbelief. Plus they announced it right after the breakup with MobileEye so the whole thing was a bit fishy. I didn't buy into until 2018, and what I was buying was the HW upgrade promise.

Where I work I tend to be the early adopter of anything new, and FSD was no different.

I can't say how many of my engineering co-workers believed Tesla would achieve Autonomous driving because I down played the entire thing from the beginning, and none of them had strong enough feelings about autonomous driving for any significant conversation to happen.

I remember demonstrating AP to a few of them with the plan of doing that again when FSD was released, but the FSD Beta is too embarrassing to show anyone.

The concern your neighbor has is likely a mix of the illusion of being in control, and as a SW guy he's all too familiar with buggy SW. Heck one of the things I like the most about autonomous driving in general is we're finally faced with a challenge where the SW must work. I think the entire SW industry has gone way too far down the "we'll fix it later" rabbit hole.

 

[Knightshade]

What's clear is what's being tested, and not what's intended.

No, that's clear in the CA DMV stuff too as far as the city streets code.

Read specifically their remarks on the OEDR systems capabilities and intent.

Regardless of what people feel they promised we all can agree on what's intended, and what's intended certainly isn't L2.

This is factually wrong.

What is being tested by the public beta right now is explicitly intended as L2

It tells you this in the docs.

Getting above that will require additional components not in the stuff being tested today

(at least by the public- it's certainly likely there's internal stuff)

 

[S4WRXTTCS]

This is a good explanation.

I think people who are angry with Tesla/Elon assume Elon was lying when he said FSD would be ready by end of the year etc. They don't seem to be able to accept (what I think is) the fact that Elon was simply wrong. Multiple times.

Infact a lot of people think he has always been completely wrong, even now and only Lidar and HD maps can solve FSD.

Ofcourse, there is a significant overlap between the above two groups (people who think world's richest man is a fraud/charlatan).

I'm firmly of the belief Tesla lied (or grossly misled buyers) about both EAP, and FSD.

With EAP I felt as if it was a necessary lie. That sometimes you're forced to lie to survive, and the only way out of the predicament they were in was to lie.

With FSD I felt as if it was an unnecessary lie. There was no need at the time to promise FSD, and they simply weren't at a point where any reasonable person would promise FSD by the end of the year.

By saying Elon was simply wrong is being really charitable to both Tesla, and to Elon.

It also dismisses the fact that a lot of Engineers quit over it.