Autonomous Car Progress

[Knightshade]

Well @verygreen finally clarified one thing for me:
View attachment 634223
Do tell green, what did they lie about, actual specifics (or are we too stupid to understand the technical nuance in your opinion?)
https://twitter.com/greentheonly/status/1352431051134672896

It kind changes your perspective when the info you read from a poster is tainted from out of the gate with some sort of vendetta or agenda (i.e. lied about AV progress)

I know you hate facts- but this isn't a vendetta, it's an accurate description of what happened.

Do you really need the long list of broken autonomy promises and deadlines rehashed to you for a 8000th time so you can again pretend you don't already know them all?


None of that is to say they won't eventually get there.

But they've sure as hell not hit a slew of previous promised target dates, going back years.

 

[mikes_fsd]

this isn't a vendetta, it's an accurate description of what happened.

Do you really need the long list of broken autonomy promises and deadlines rehashed to you for a 8000th time so you can again pretend you don't already know them all?

Fact 1: He bought a $100k car (Model X) in his words "I got a Tesla because they lied about AV progress"
Fact 2: And his first ever Tesla video is of Autopilot in debug mode

( May 14, 2017 ) almost exclusively posting about Autopilot since.

That sure looks like he was trying to prove more than just deadlines missed, he is/was trying to prove that the Tesla hardware and software are not up to the task.

Simple follow up question: Four years of trying to show how "they lied about AV progress" and no smoking gun?
Okay then!

 

[daniel]

Never buy a thing for what someone promises it will be able to do "someday." Buy it for what it will do when you get it, if that's something you want. I bought my Model 3 because it's the best car on the road with features I wanted. I didn't pay for FSD, but I will when and if they get there in my lifetime.

Referring (I believe) to V2X:

how would you implement for a 20% adoption rate?

I wouldn't implement it because I'm not an engineer and wouldn't be capable of it. Back when op amps were a thing, I never managed to get one to actually do anything. But I'm confident that the engineers at car and AV companies could work out protocols that would allow cars to get some benefit from being able to communicate with a small number of other cars, getting better and better as more cars have it.

 

[Knightshade]

Fact 1: He bought a $100k car (Model X) in his words "I got a Tesla because they lied about AV progress"

And he's right.

Just a small sampling of claims about AV progress that didn't turn out to be true:



Elon said 2 years to full autonomy back in 2015. Didn't turn out to be true.

Elon said in 2016 that HW2.0 was sufficient for full autonomy. Didn't turn out to be true. (then later said 2.5 would be-also untrue)

Tesla showed an FSD video in 2016 as a single continuous demo, which it later turned out was mostly faked by splicing together over 100 different failed FSD attempt drives to look like a single successful one.

Elon said EAP and FSD would diverge on features by early-mid 2017. Didn't turn out to be true.

Elon said Tesla would do an autonomous coast to coast drive by end of 2017. Didn't turn out to be true.

Elon said everyone with FSD would be able to do such an autonomous drive by end of 2019. Didn't turn out to be true.

Elon said he was confident Tesla would have robotaxis approved in one or more jurisdictions by end of 2020. Didn't turn out to be true.

Fact 2: And his first ever Tesla video is of Autopilot in debug mode ( May 14, 2017 ) almost exclusively posting about Autopilot since.

If you think he posts "almost exclusively" about Autopilot you either haven't actually read most of his posts, or you're simply not interested in honest discussion.

Most evidence suggests that second one, but hey could be both I guess.

He posts about a lot of behind the scenes stuff, including tons unrelated to autonomy.

That sure looks like he was trying to prove more than just deadlines missed, he is/was trying to prove that the Tesla hardware and software are not up to the task.

So far, he's right. Especially if discussing the 2017 hardware that came in his X.

Tesla admits he's right in fact.

They've updated the sensor suite AND the driving computer since he got that car.

They've admitted the frame-by-frame software path they'd been doing down for several years wasn't going to reach autonomy and went to a fundamental re-write since he got that car.

Simple follow up question: Four years of trying to show how "they lied about AV progress" and no smoking gun?
Okay then!

.... you mean besides all the claims they made they ended up not being true because the things they promised by a specific date didn't actually happen?

And besides the fact they ended up having to toss out the 2.x code as incapable of autonomy after years of making claims it was?

And besides the fact they've had to make several hardware updates because the original stuff was insufficient for L5?



Sounds like literally nothing anyone could show you would count as evidence to you that Tesla was ever wrong about anything.

 

[mikes_fsd]

If you think he posts "almost exclusively" about Autopilot you either haven't actually read most of his posts, or you're simply not interested in honest discussion.

I was specifically talking about his YouTube page - not his twitter page.

Sounds like literally nothing anyone could show you would count as evidence to you that Tesla was ever wrong about anything.

I think Tesla was wrong with how they handled MCU1 failures -- waiting until they were forced to do recall.

As for Autopilot - people want to use "but Elon" and refuse to use what Tesla officially published - Autopilot - neither in 2016 nor in 2021 did it give concrete timelines.

When they hit an impasse and needed to upgrade the computer, they did on their own initiative.

As - Autopilot - stipulated features have been released by OTA and I've seen and see released - building blocks - and I see steady improvement in the software suite.
What Tesla has shown in this time is that they are working the problem and if they find a blocker (like the need for HW3 FSD computer) they solve the problem by addressing it, not backing away from it.

 

[Knightshade]

I was specifically talking about his YouTube page - not his twitter page.

Then you used the word "exclusively" incorrectly.

I think Tesla was wrong with how they handled MCU1 failures -- waiting until they were forced to do recall.

Agreed.

As for Autopilot - people want to use "but Elon" and refuse to use what Tesla officially published - Autopilot - neither in 2016 nor in 2021 did it give concrete timelines.

A few things here.

1) He is the CEO (and also chairman during most of those statements). His statements are material when speaking as to the companies progress on things. Just ask the SEC.


2) Tesla, not just Elon, made several of the turns-out-not-actually-true claims I cited in my post on both hardware and software capabilities. INCLUDING having given TIMELINES of certain features they failed to deliver as promised.

For example Tesla (not just Elon) promised Stop sign/light stopping, and autosteer on city streets, by end of 2019.

Neither was delivered as promised.

Lights/signs finally came in 2020, with autosteer in the city promise now replaced with a by end of 2020 promise.

Which they also broke (unless you're among the 1000 early beta FSD testers I suppose)


They also, as another example, lied about the capabilities of Smart Summon for at least months on their website, claiming it'd come get you anywhere in a parking lot (rather than the short range, line of sight, capabilities it actually offers)--- I was actually shocked how long they left that wording up there given how untrue is was for a feature they were actively selling.

When they hit an impasse and needed to upgrade the computer, they did on their own initiative.

YM
to avoid all the lawsuits because they lied about the original HW being capable of it
HTH!

As - Autopilot - stipulated features have been released by OTA and I've seen and see released - building blocks - and I see steady improvement in the software suite.
What Tesla has shown in this time is that they are working the problem and if they find a blocker (like the need for HW3 FSD computer) they solve the problem by addressing it, not backing away from it.

Absolutely.

I said several posts ago I believe they're still intending to reach all the goals they've laid out.

And absolutely applaud their willingness to realize one path is a failure and pursue another one.


That does not change the fact they've made a lot of specific claims about the pace of progress, and target dates for various abilities and features, that turned out not to be true.

 

[mikes_fsd]

That does not change the fact they've made a lot of specific claims about the pace of progress

And I've said elsewhere on the forum that I understand arguments/complaints about missed deadlines (it was in the context of not understanding ppl that never bought FSD and still complain about FSD ad nauseum on here)

For some reason, I don't feel wronged/misled/hoodwinked and I bought the same FSD as the rest of Tesla FSD purchasers... the only way I can reconcile that is ppl did not want to take company stated timelines (i.e. no real timeline) listed here Autopilot -- but want to take the "but Elon" approach instead.

 

[verygreen]

Do tell green, what did they lie about, actual specifics (or are we too stupid to understand the technical nuance in your opinion?)

in case you missed it, when I bought in Jan 2017 this is what was on sale (and similarly puffed up "FSD" that I also purchased).

Needless to say it was a lie and they were not only not finishing any validation, they did not even start writing most of that code.

This contributed a great deal to my desire to find out the REAL state of progress once I realized the company lies about it. There's no vendetta as you put it. I just seek the truth (or as close to it as I may get anyway given that I don't have inside sources)

 

[Bladerskb]

As for Autopilot - people want to use "but Elon" and refuse to use what Tesla officially published - Autopilot - neither in 2016 nor in 2021 did it give concrete timelines.

The AP page and order page literally gave a timeline of December 2016 for EAP features. They didn't meet it, so they changed the verbiage afterwards. EAP ended up coming Oct 2018. Two years later. I could go on and on..

 

[daniel]

I've said it before: Elon is a "chrono-optimist." He thinks everything can be done much sooner than it actually can. If you understand this you're good. You'll take his timeline promises with a fistful of salt. The problem is that many in the general public see what he's accomplished and take him at his word. They pay for a feature that he's promising in a year or two, because it costs less if you buy it with the car, and when they still don't have the feature a decade later they're upset. And they have every right to be.

I'm not upset because I only paid for what was available at the time and it works like a dream. But I love Tesla and I admire Elon and I own a few shares of the stock, and it disappoints me when he makes promises he cannot keep. He could have left it at "We're working on it and we think we can do it," but instead he made concrete promises that were not necessary because these cars sell themselves. And so a lot of people are upset that they paid for something they never got and probably will not get until their car is old and it's time to trade it in.

Here's what he should have done, and I'd have taken him up on it: When you buy the car you have the option of buying an FSD bond: For $5,000 or $8,000 or $10,000 or whatever the amount is, you get a ten-year bond that pays 5% interest. (That's what Tesla pays me for my solar bonds I bought from Solar City.) If Tesla provides your car with driverless FSD within the ten years, Tesla gets the bond as payment. If at the end of ten years Tesla has not upgraded your car, either because they were unable or because you no longer own the car, the bond matures and you get your principle back. You've invested a few thousand dollars at a decent rate of interest. If you sell your car you have the option of keeping the bond and collecting the interest until maturity (the buyer might not want FSD) or transferring the bond to the new owner. This would have been fair to the buyers.

 

[S4WRXTTCS]

in case you missed it, when I bought in Jan 2017 this is what was on sale (and similarly puffed up "FSD" that I also purchased).

Needless to say it was a lie and they were not only not finishing any validation, they did not even start writing most of that code.

This contributed a great deal to my desire to find out the REAL state of progress once I realized the company lies about it. There's no vendetta as you put it. I just seek the truth (or as close to it as I may get anyway given that I don't have inside sources)

Fundamentally speaking the entire EAP/FSD package was born from two lies.

The first lie being that EAP was anywhere close to matching the feature set or performance of AP. It took so long that Tesla ended up having to settle a lawsuit where the earlier buyers got a little money from EAP being so bad.

On that one I don't fault Tesla for the lie as they were between a rock, and hard place. Basically MobileEye fired them as a customer when they caught Tesla trying to put HW2 + MobileEye into a car at the same time. Once that happened they basically had to either tell customers that they'd have to wait for AP features or they'd have to lie. They chose to lie.

What I do fault Tesla for is the FSD lie. I refer to that one as the unnecessary lie. We know its a lie because no one can promise FSD. No one has accomplished FSD to the extent that it was marketed as Tesla did. Quite a bit of what it will take to reach L4/L5 autonomous driving is outside of Tesla's hands. Sure they can guess at what it will take, but you can't sell a product based on a guess.

We also know that Tesla considerably watered down what FSD is on order page. I can't say I'd get all that excited about AutoSteer in city streets. If I had purchased when that change happened I likely wouldn't have purchased it.

I'm not sure how an FSD owner from 2017 is supposed to think that it was anything, but a lie. Sometimes people simply don't mind being lied if they feel like it contributed to the greater good. I understand the logic as in the fact that the autonomous driving promises likely contributed a fair amount to the media coverage of Tesla. Like right now the best marketing Tesla has is the FSD Beta Youtube videos, but its not clear when that will ever reach our cars.

 

[daniel]

Fundamentally speaking the entire EAP/FSD package was born from two lies.

The first lie being that EAP was anywhere close to matching the feature set or performance of AP. It took so long that Tesla ended up having to settle a lawsuit where the earlier buyers got a little money from EAP being so bad.

On that one I don't fault Tesla for the lie as they were between a rock, and hard place. Basically MobileEye fired them as a customer when they caught Tesla trying to put HW2 + MobileEye into a car at the same time. Once that happened they basically had to either tell customers that they'd have to wait for AP features or they'd have to lie. They chose to lie.

What I do fault Tesla for is the FSD lie. I refer to that one as the unnecessary lie. We know its a lie because no one can promise FSD. No one has accomplished FSD to the extent that it was marketed as Tesla did. Quite a bit of what it will take to reach L4/L5 autonomous driving is outside of Tesla's hands. Sure they can guess at what it will take, but you can't sell a product based on a guess.

We also know that Tesla considerably watered down what FSD is on order page. I can't say I'd get all that excited about AutoSteer in city streets. If I had purchased when that change happened I likely wouldn't have purchased it.

I'm not sure how an FSD owner from 2017 is supposed to think that it was anything, but a lie. Sometimes people simply don't mind being lied if they feel like it contributed to the greater good. I understand the logic as in the fact that the autonomous driving promises likely contributed a fair amount to the media coverage of Tesla. Like right now the best marketing Tesla has is the FSD Beta Youtube videos, but its not clear when that will ever reach our cars.

^ This!

Some FSD buyers have said they are happy to pay Tesla the money for FSD so that Tesla can develop it. I would gladly buy Tesla bonds to help them fund the research. But just giving them money is another thing. I own some shares and I would buy some bonds. But giving money to a for-profit corporation is another matter entirely.

I love EAP, but I bought it when it was an existing feature and I knew exactly what I was buying and it worked pretty much as advertised, and now with improvements it does everything I expected of it. ( Well, except for self-park, and summon which I've never even tried because I have no need of it and was not part of my reason for buying EAP.)

Tesla said that my HW 2.5 car had all the necessary hardware to become a driverless robotaxi. They said this at a time when they could not possibly have known that because they didn't have (and still don't have) the software. The companies that do have robotaxis (within very limited and well-mapped geographical areas) have significantly more sensors than Tesla, yet Tesla still says that its (current) hardware is sufficient, without having the software. HOW CAN YOU KNOW FOR CERTAIN WHAT HARDWARE IS NEEDED FOR SOFTWARE THAT DOES NOT YET EXIST?

 

[NHK X]

Never buy a thing for what someone promises it will be able to do "someday." Buy it for what it will do when you get it, if that's something you want. I bought my Model 3 because it's the best car on the road with features I wanted. I didn't pay for FSD, but I will when and if they get there in my lifetime.

Referring (I believe) to V2X:


I wouldn't implement it because I'm not an engineer and wouldn't be capable of it. Back when op amps were a thing, I never managed to get one to actually do anything. But I'm confident that the engineers at car and AV companies could work out protocols that would allow cars to get some benefit from being able to communicate with a small number of other cars, getting better and better as more cars have it.

Well since you made a claim that it could benefit even at early adoption rates thought that you had some sort of idea or read something about it. Not expecting an engineers answer. So what you are saying is that you have no idea but threw that out there.

I got nothing against V2x, think the idea is radical. And genuinely thinking would be awesome to see. I just have a hard time seeing how it could work with a 20% adoption rate.

 

[S4WRXTTCS]

I love EAP, but I bought it when it was an existing feature and I knew exactly what I was buying and it worked pretty much as advertised, and now with improvements it does everything I expected of it. ( Well, except for self-park, and summon which I've never even tried because I have no need of it and was not part of my reason for buying EAP.)

I bought in late 2018, and I traded in a perfectly good Model S 70D. I got EAP+FSD because I felt like EAP would never really be that good without HW3. I bought at a time when HW3 was being promised for buyers of FSD. I got FSD despite hating the way it was sold. I really wish they would have had basic AP separated from FSD back then. Where FSD was simply "free autonomous SW/HW upgrades" for X years. Where they simply had a roadmap, and a bug tracking site. So basically you just submitted bugs, and problematic location for Tesla to fix for various L2 level capabilities (NoA, smart summon, etc). It still would have pushed the industry while not being an embarrassment.

FSD has ruined so much of the car

The awkwardness of trying to get a rational friend to ignore the whole FSD thing, and focus on just the car.
No driver monitoring as it wouldn't be needed for FSD. By far this biggest annoyance for myself while driving with AP is the nag. The Model S I had was so much better as the torque sensor was more sensitive, and the algorithm simpler.
No ability to transfer FSD to a new Tesla so you're stuck in the one you have.
No publicly announced valuation for FSD on a trade in. So you never really know how much you've been screwed by

Most of all FSD has ruined EAP. You love EAP, but even you would likely admit that its inconsistent. As an engineer I hate inconsistency a lot. To me inconsistency is a failure, and every feature in EAP is a failure due to inconsistency.

it's not just .1% inconsistent, but we're talking on the order of 20% or more.

Why is it so bad?

There is no feedback loop for owners to report problems in a fashion where we're told that they're being fixed or worked on. As an example I can report a maps issue to Apple, and it will get fixed promptly. They'll thank me for my contribution so that motivates me to report things in the future.

There doesn't seem to be any effort at fixing problems, and instead all the effort is on FSD. The FSD beta people have the report button.

The inability for Tesla to rev the hardware to address known issues. As a result Tesla/Elon have a "fix it in SW" mentality, and quite honestly SW fixes tend to be inconsistent. Where at best they're a patch.

As a result of the bet on FSD means that everything is riding on the FSD beta.

If Tesla releases the FSD beta to the general public in the next 12 months I think a lot of things will be forgiven. The reason is its so fundamentally different than the current FSD+EAP that a lot of the inconsistencies will go down. If they leave the report button than a lot of the "this issue has been there forever" problems will hopefully get addressed.

If Tesla releases Autosteer on city streets I expect a huge backlash. That might be what's on the order page, but that isn't what people are expecting. You can't tease hundreds of thousands of owners with FSD beta videos, and then turn around and limit it to certain things.

Once FSD beta is on the roads I think it meet the "you've shown a credible effort" requirement that buyers naturally have.

The other thing that will change everything is the subscription model. The subscription model will push resale values up of FSD equipped Tesla's. I'm generally pretty happy if I bought, and sold something for the same price even if it never matched what I was hoping from it. Like if Tesla gave me $8K for EAP+FSD I wouldn't hold it against them.

 

[S4WRXTTCS]

Well since you made a claim that it could benefit even at early adoption rates thought that you had some sort of idea or read something about it. Not expecting an engineers answer. So what you are saying is that you have no idea but threw that out there.

I got nothing against V2x, think the idea is radical. And genuinely thinking would be awesome to see. I just have a hard time seeing how it could work with a 20% adoption rate.

The problem with V2x is its a solution to a problem that doesn't currently exist.

It's like planning a city for cars when cars haven't been invented yet. But, if you don't account for it then you've designed a city where driving is a nightmare.

I think eventually it will happen in some form or another. Sometimes things just take a bit to happen.

Years ago I tried to get approval of tele-presence robots at my place of employment. It languished, and the one robot we got was stuck in a meeting room as EHS people didn't want it hurting anyone . They were worried that this object would move, and people might trip over it. In any case it was QUICKLY approved during covid for full movement throughout the floor its on. Not just approved, but we ordered another one.

 

[diplomat33]

Tesla said that my HW 2.5 car had all the necessary hardware to become a driverless robotaxi. They said this at a time when they could not possibly have known that because they didn't have (and still don't have) the software. The companies that do have robotaxis (within very limited and well-mapped geographical areas) have significantly more sensors than Tesla, yet Tesla still says that its (current) hardware is sufficient, without having the software. HOW CAN YOU KNOW FOR CERTAIN WHAT HARDWARE IS NEEDED FOR SOFTWARE THAT DOES NOT YET EXIST?

Elon believes his FSD approach will work so he just treats it as fact that the hardware is "FSD capable". Basically, Elon is starting with the assumption that the hardware is good enough for FSD and that it is just a matter of finishing the software.

Now, I don't fault Elon for believing in his FSD approach. But I do fault him for making false statements about FSD and for selling FSD before it exists. You can't just declare the hardware to be "FSD capable" (and sell non-existent FSD to customers) before you actually have FSD, just because you believe the approach will work in the future, which is what Elon is doing.

Other AV companies know that accurate and reliable perception is essential. And they want to minimize perception failures as much as possible. So they start with trying to give the AV the best perception hardware possible. They use cameras but they also include other sensors like HD maps, lidar and radar to help make the perception as complete, accurate and reliable as possible. It's not a problem if you have too much data. It is a problem if you have too little data. For example, if your AV does not detect another vehicle because there was no camera covering that angle or the camera data was inconclusive, that's a fatal flaw. Basically, they err on the side of maybe too many sensors to make sure the car's perception is good enough.

 

[NHK X]

Never buy a thing for what someone promises it will be able to do "someday." Buy it for what it will do when you get it, if that's something you want. I bought my Model 3 because it's the best car on the road with features I wanted. I didn't pay for FSD, but I will when and if they get there in my lifetime.

Referring (I believe) to V2X:


I wouldn't implement it because I'm not an engineer and wouldn't be capable of it. Back when op amps were a thing, I never managed to get one to actually do anything. But I'm confident that the engineers at car and AV companies could work out protocols that would allow cars to get some benefit from being able to communicate with a small number of other cars, getting better and better as more cars have it.

And honestly, the arguments you made in each paragraph are a little antithetical. In one you argue against expecting some benefit now but hoping for much larger benefits later, and the other you argue for the idea that far out tech can have some benefits now (without any solid example of how a intersection could do it now).

 

[daniel]

And honestly, the arguments you made in each paragraph are a little antithetical. In one you argue against expecting some benefit now but hoping for much larger benefits later, and the other you argue for the idea that far out tech can have some benefits now (without any solid example of how a intersection could do it now).

Something as complex as FSD or V2X will not come in one fell swoop with the flick of a switch. Eventually we will have both. For both, we should expect gradual progress and gradual improvement to completion.

My argument for V2x is that even though there will not be 100% deployment with the flick of a switch, the gradual implementation will be useful because more data is always good, as long as you have the processing power to handle it. Saying that my argument is invalid because I don't have the technical details (as one person did above) is like a certain discredited politician telling Greta Thunberg that she should shut up about global warming because she doesn't have the technical answers.

My argument about FSD is not that we won't get it or that Tesla should stop making incremental improvements. My argument is that Tesla should not have promised people that their HW 2.0 and later HW 2.5 cars had "all the necessary hardware" for FSD when they didn't have the software and couldn't possibly have known what hardware would be needed.

I did not say that FSD buyers should not expect some benefit now! I did NOT say anything of the sort. I said that Tesla should not have promised buyers that their car, with the hardware in it at the time of purchase, would be able to drive itself across town to pick the kids up from soccer practice and bring them home with no driver in the car. And I said that Tesla should not have promised that for $8,000 (or whatever the amount was) their car would do those things when Tesla had no way to know whether the cars then being sold would be able to do that.

 

[NHK X]

Something as complex as FSD or V2X will not come in one fell swoop with the flick of a switch. Eventually we will have both. For both, we should expect gradual progress and gradual improvement to completion.

My argument for V2x is that even though there will not be 100% deployment with the flick of a switch, the gradual implementation will be useful because more data is always good, as long as you have the processing power to handle it. Saying that my argument is invalid because I don't have the technical details (as one person did above) is like a certain discredited politician telling Greta Thunberg that she should shut up about global warming because she doesn't have the technical answers.

My argument about FSD is not that we won't get it or that Tesla should stop making incremental improvements. My argument is that Tesla should not have promised people that their HW 2.0 and later HW 2.5 cars had "all the necessary hardware" for FSD when they didn't have the software and couldn't possibly have known what hardware would be needed.

I did not say that FSD buyers should not expect some benefit now! I did NOT say anything of the sort. I said that Tesla should not have promised buyers that their car, with the hardware in it at the time of purchase, would be able to drive itself across town to pick the kids up from soccer practice and bring them home with no driver in the car. And I said that Tesla should not have promised that for $8,000 (or whatever the amount was) their car would do those things when Tesla had no way to know whether the cars then being sold would be able to do that.

I still don’t see how you can gradually implement v2x. It is not an issue of data.

 

[drtimhill]

ok everyone, so hindsight is 20/20 .. Tesla made promises and didn't keep them. Some people took those promises at face value, and got burned. Shouldn't have happened, but it did. I took a risk with FSD, but I went in knowing they would maybe deliver 50% max on the promises, so I'm perhaps not as angry as others.

Now what? Sure, we should not forget the broken promises ("Those who forget history..." etc). But just saying again and again "I got burned, I'm angry" isnt going to achieve anything. Time to move on, people.

At this point in time, Tesla are much closer than anyone to getting a car to drive itself on city streets and freeways with minimal intervention. No-one else is close. Nope, not MobileEye, not GM cruise, not Waymo. Sure, they have cool looking videos and slide decks, but so did Tesla. You think their promises are any less nebulous than Tesla's from 3 years ago?

Right now, the other guys are where Tesla was years ago, making promises and slick videos. Tesla has stuff actually out there in the hands of ordinary drivers. I know where I'd bet my money .. it's where I bet it when I paid for FSD. Maybe I'l be wrong, maybe I'll be right, but it's a bet, right? It's called speculation, if you don't like it, buy a Toyota and enjoy it, they are great cars, and last forever.