Seems like FSD is a complete crock

[scott1002]

Maybe. In which case you should change what you want and buy a Tesla anyway. Sadly, you'll have to deal with a car that's only the best thing you've ever driven. And it will gradually morph into the perfect thing you want.

This is a far superior experience to driving something else until your dream car appears.

The 2020 mid engine corvette is the other option under consideration. That seems pretty perfect too. Except for the interior; seems a bit ugly to me.

 

[Bet TSLA]

The 2020 mid engine corvette is the other option under consideration. That seems pretty perfect too. Except for the interior; seems a bit ugly to me.

Never driven one. Do you like how it drives better than a Tesla?

 

[EVNow]

That seems pretty perfect too.

Not sure how anything that spews toxins into the air can be "pretty perfect"

 

[iqless]

After initially disliking it, I'm actually glad Tesla made the change to include AP and isolate the FSD features - other cars have some (inferior) version of AP, so that's no longer a major differentiator.

So as the functionality of FSD evolves, it will all be within that bucket of capability.

Two problems-- one, Tesla has been quick to make promises on capabilities but delayed on keeping them. Two, our society has become accustomed to getting what we want RIGHT NOW. Hell, we order stuff on Amazon and it gets here the same day.

If Amazon told me I'd get a package between 3 and 6 months and I still didn't have it after 2.5 years I'd have a problem with that and it's not because I'm a lazy entitled impatient millennial

 

[Octo]

Disagree. He did give a positive review for AP and NOA, but the keyword is that he said it is HIS experience. Others may have negative experiences depending on various factors. I don't think he was shaming anyone or blanketing a glowing review of AP/FSD for everyone.

Well, he did tell me that I should get a car from 1960 because I told him I think that Tesla’s autonomy features are useless to ME because of the limited reliability (Personally I think it’s a cool tech demo but with the current release, it’s not really useful because *nothing* is guaranteed and the driver MUST be alert all the time. I mean, what’s the point, if effectively I need to supervise the equivalent of the most idiotic student driver ever).

As if these features are the main value of a Tesla.

That put him into fanboy category, IMO.

 

[Saghost]

For that to happen, first Tesla has to deliver all the features (see my FC thread). Then they need a billion or so miles driven by the fleet on FSD/AP. Then, they can assess how many accidents they would have had if the driver didn't intervene. Only then they can figure out what the risk/liability is and take ownership of any problems. That is some years away.

Disagree. There's no reason the car has to jump from level 2 straight to doing level 3 everywhere in one step.

There's a lot of opportunity for level 3/4 that's limited to easier environments, and you don't need to be feature complete for all environments for that - you need all the features that relate to that environment, but not ones that don't.

To me the obvious path is to get obstacle detection and avoidance working and then work towards certifying FSD for the freeway environment, while they're still working on the stoplight and stop sign and all the other challenges of the urban environment.

A car that truly drives itself on the freeway in good weather has a lot of value for people and is much closer to ready than a car that drives itself everywhere. (Then next year it can drive country roads, then suburbs, and finally the city environments and bad weather.)

 

[scott1002]

Never driven one. Do you like how it drives better than a Tesla?

No one has driven one. And my dealers said they’ll all be presold with zero opportunities to test drive.

No 7 day returns either!

 

[Bet TSLA]

No one has driven one. And my dealers said they’ll all be presold with zero opportunities to test drive.

No 7 day returns either!

Okay then. Some people get way more pleasure living in fantasy land where everything is unicorns and rainbows. Others, like me, prefer to have what I'm doing be as wonderful as it can be. I own two Teslas, a Model 3 and a Model S and love driving them. Best purchases I've ever made. I enjoy a good fantasy too, but reality wins every time.

You should do what works for you. Fantasy has the huge advantages that it's way cheaper and hassles are non-existent.

 

[jebinc]

No one has driven one. And my dealers said they’ll all be presold with zero opportunities to test drive.

No 7 day returns either!

You have to return in before 7 days or 1000 miles... then wait a fairly long period for the refund to arrive.

 

[scott1002]

If Amazon told me I'd get a package between 3 and 6 months and I still didn't have it after 2.5 years I'd have a problem with that and it's not because I'm a lazy entitled impatient millennial

I work for Amazon. It’s a big problem if it’s a day late. As it should be. Tesla take note.

 

[iqless]

You have to return in before 7 days or 1000 miles... then wait a fairly long period for the refund to arrive.

60 days to be exact. They'll hold on to your $40k-120k for 60 days.

Tesla customers describe maddening problems with returns and refunds

 

[scott1002]

Okay then. Some people get way more pleasure living in fantasy land where everything is unicorns and rainbows. Others, like me, prefer to have what I'm doing be as wonderful as it can be. I own two Teslas, a Model 3 and a Model S and love driving them. Best purchases I've ever made. I enjoy a good fantasy too, but reality wins every time.

You should do what works for you. Fantasy has the huge advantages that it's way cheaper and hassles are non-existent.

Haha. It’s not clear to me what side is the fantasy and unicorn side. I am much more confident the corvette will operate as anticipated from day one.

The tesla is a much more practical day to day car (I think). I’m reluctant to drive my Porsche in certain situations. Too many my wife would say. I think the corvette would be like that so the plan for the Tesla is to be the everyday car.

 

[SSonnentag]

“Well, we already have Full Self-Driving capability on highways. So from highway on-ramp to highway exiting, including passing cars and going from one highway interchange to another, Full Self-Driving capability is there. In a few weeks, we’ll be pushing an update that will allow the option of removing stalk confirm (for Navigate on Autopilot) in markets where regulators will approve it, which would be the case in the US, for example."
Elon Musk

It's definitely beta though!

Perhaps Elon already has functional FSD on his car, but none of us do. The example from within the past week illustrates how FSD on freeways, let alone highways, is not yet here. The fact that we have to keep letting the car know we're still alive ever 20-30 seconds is further proof of non-existent FSD.

 

[Daniel in SD]

Perhaps Elon already has functional FSD on his car, but none of us do. The example from within the past week illustrates how FSD on freeways, let alone highways, is not yet here. The fact that we have to keep letting the car know we're still alive ever 20-30 seconds is further proof of non-existent FSD.

Yeah, it does seem like running into large objects is the bug they should be focusing their efforts on right now. I'm not a software engineer though.

 

[socal_suburbanit]

Here is why I think we are still years away from FSD-the many departures from Tesla in the last 18 months. Remember that senior positions hold thousands of shares of stock and Tesla would explode in value if they actually delivered just what Elon has predicted. (As I write this we still await Enhanced summon which he said was weeks away back in November 2018.)

CTO Straubel leaves in 7/2019 (will now be in an advisory role)
CFO Ahuja leaves in 1/2019
GC Maron announced departure 12/2018
VP Engineering Field 8/2018
VP Sales and Service McNeill 2/2018
HR chief Toledano 9/2018
Chief Accounting Officer Morton 9/2018
Head of Digital Product 12/2018
Treasurer and VP Finance Repo 3/2018

Autopilot perception lead Popovic 6/2019
Musk restructures team to reduce Bower's responsibilities and have more report to him 5/2019
Roughly 10% of the Autopilot software team have left in the last few months.

Autopilot is an extremely difficult thing to create in software. We can see from the outside turnover and chaos in the AP team as well as many senior people who would sit in on high level status meetings deciding to leave the company pretty much spells out that Tesla are not close to meeting their goals. I watched Autonomy day and was really impressed by the NN presentations but seeing the churn of the teams since then has been cold water on my expectations.

When I first bought my M3 and started paying attention to the news I thought Elon was just enthusiastic and optimistic about his announcements of what is coming soon with Tesla. But the fact he repeatedly is orders of magnitudes off on his timelines about the same topic (again, take Enhanced Summon as an example) shows that he is not intellectually honest about what Tesla is able to deliver. I can only assume he is in 'fake it until you make it' mode in an attempt to get cash flow through the company. Honestly I am surprised the SEC hasn't done more.

And consider how long do you think the jump to city street AP even with supervision will be in early release? Seems like 6-12 months would be reasonable. I expect a sale on FSD each quarter end until they get to early release to try to generate cash.

To your question, is FSD a complete crock? No, I don't think that is fair. But are the FSD timelines Elon discusses a complete crock? I think the answer to that is sadly, yes.

 

[EVNow]

I watched Autonomy day and was really impressed by the NN presentations but seeing the churn of the teams since then has been cold water on my expectations.

What churn ? Karpathy is still there. If he leaves, I'll be worried. Churn is constant in silicon valley - in all companies, in all teams. BTW, don't assume all of the people "left". I'm sure a lot of them were let go / asked to go.

When I first bought my M3 and started paying attention to the news I thought Elon was just enthusiastic and optimistic about his announcements of what is coming soon with Tesla. But the fact he repeatedly is orders of magnitudes off on his timelines about the same topic (again, take Enhanced Summon as an example) shows that he is not intellectually honest about what Tesla is able to deliver. I can only assume he is in 'fake it until you make it' mode in an attempt to get cash flow through the company. Honestly I am surprised the SEC hasn't done more.

Because - he is not "faking" it. The way Musk operates is - he thinks of only happy outcomes and happy paths. That is why he is optimistic on timelines.

At the same time, that is also the reason he is a visionary. Do you think Tim Cook would have started SpaceX or an EV startup in an industry that hasn't had a successful new company in 100 years ?

 

[diplomat33]

But the fact he repeatedly is orders of magnitudes off on his timelines about the same topic (again, take Enhanced Summon as an example) shows that he is not intellectually honest about what Tesla is able to deliver. I can only assume he is in 'fake it until you make it' mode in an attempt to get cash flow through the company. Honestly I am surprised the SEC hasn't done more.

I don't think he is "faking it". That would assume that Tesla does not have enhanced summon or whatever feature he is promising. That is clearly not the case. We've seen videos that prove enhanced summon is real.

Maybe Elon is just really bad at making realistic deadlines? I think that's a simpler and more logical explanation.

 

[Octo]

That is clearly not the case. We've seen videos that prove enhanced summon is real.

Maybe Elon is just really bad at making realistic deadlines? I think that's a simpler and more logical explanation.

So what about the 2016 FSD video? Do you also think Tesla keeps fully working FSD in their secret labs and only releases AP that can’t detect traffic cones/barrels to the public for added suspense?

Because, we have seen a video of a Tesla car three years ago that was driving itself.

If it’s on video “it must be real”. Bonus realness, if it is a video on the internet.

 

[diplomat33]

So what about the 2016 FSD video? Do you also think Tesla keeps fully working FSD in their secret labs and only releases AP that can’t detect traffic cones/barrels to the public for added suspense?

Because, we have seen a video of a Tesla car three years ago that was driving itself.

If it’s on video “it must be real”. Bonus realness, if it is a video on the internet.

We've discussed the 2016 FSD video ad nausea on this forum. Basically, as far as I can tell, the video was real but it was just a demo, not a finished FSD. In simple terms, the timeline is basically the following: after the mobileye break up, Tesla had to start AP from scratch. They had some success enough to string together a demo to show what FSD would look like when finished and to promote the AP2 hardware. But then Tesla struggled to make FSD work reliably. So they decided to start from scratch again. Tesla turned to Karpathy and his machine learning and neural net approach which has produced some good results as evidenced by the EAP and NOA features that we got. Now, Tesla hopes that this approach will continue to work all the way to a finished FSD.

 

[EVNow]

So they decided to start from scratch again. Tesla turned to Karpathy and his machine learning and neural net approach which has produced some good results as evidenced by the EAP and NOA features that we got. Now, Tesla hopes that this approach will continue to work all the way to a finished FSD.

Karpathy started in Tesla in Jun '17. So, in reality we can say Tesla's FSD program is 2 years old. They have made a lot of progress compared to so many others in the industry.

ps : Was looking at Karpathy on LinkedIn. Apparently he is "3rd" from me - so I know someone, who knows someone connected to Karpathy. Just have to figure out who