Frustrated with FSD timeline

[Vitold]

So on 23rd Jan 2017 Elon promised that FSD feature will be here. Its 9 days from 3 months mark and we are not even at full parity with AP1.

Let alone EAP specific features such as autosteer+ on non-divided local roads, on ramp / off ramp, self lane changing, smart summon and self parking.

We are supposed to have that weeks ago and preparing to get the update to enable something like stopping at traffic lights, stop sign, make turns, navigating roundabouts maybe?

Right now we have nothing, nothing but more hype statements to sell more cars, raise stock market and have naïve people pay $8,000 of their hard earned money on a fantasy FSD.

Its been 7 months and we still don't have full parity AP1 after being promised we would have not only it but EAP features.

The "6 months definitely" doesn't matter because the 3 months was used to generate hype and headlines when tesla isn't even anywhere close.

Its like saying we will beat the Warriors in 4 games but 7 games definitely.
Making that statement you already know what everyone pays attention to is that you said you will sweep the warriors and that will be the headline and the point of the statement.

I predicted exactly how long the updates would take and was down voted to hell and yet I was right.
Enhanced Autopilot 2 may never happen before Full Self Driving is ready!



Elon being challenged about his 3 months statement points to his "maybe"

Tesla enabled the second camera.

 

[lunitiks]

Yep, MobilEye and Tesla broke up in july '16
TESLARATI.com

 

[electracity]

Regardless of legality, what was Musk's motivation for taking FSD money and suggesting the functionality was in the near future?

This situation is like the model X where Tesla held a fake launch. Tesla doesn't do fake product launches because Musk is optimistic. They do fake product launches because they feel they have to.

I assume they fake launched new autopilot because it was unacceptable to admit that they were taking a functional step backwards from AP1.

Fake is the new reality. A lot of press still reports Tesla as the autonomous driving leader. So I guess fake works. The model X turned out good, so maybe the new AP will be great.

The model 3 launch will be fake too, but that is another topic.

 

[AnxietyRanger]

Tesla enabled the second camera.

EAP was supposed to use four cameras in December 2016 according to the Design Studio... so two won't help with FSD much.

5+ cameras, now, that would be better...

 

[gowthamn]

I thought by now people would have reloaded not to take Elon's words at face value. He is largely wrong with both Tesla and SpaceX.

 

[McRat]

I feel like I just watched a movie on the Lifetime channel.

I had to look that up. The fact you know there is a Lifetime Channel is perfectly OK. Diversity rocks!

But the point is, MOST software ships on time because more and more resources get thrown at it if the timetable starts to slip.

 

[Naonak]

Ugh... that really sucks that AP2 stuff is late, but it's not surprising.

What I find a little more curious is that people think they are going to get what was promised with AP2. They won't - they will get what was promised with AP1 on AP2 hardware. AP2 promises will come with AP3.

AP1 fell far short of all the initial promises. Those promises look like they can be fulfilled with AP2 hardware - so we'll see true on-ramp to off-ramp driving, we'll possibly see summon (not the fake summon we have with AP1) where the car will drive to you on private roads/parking lot, stop signs & stop lights.

What you will never see on AP2 hardware is full autonomous driving where you do not have to pay attention to the road. You will never see fully reliable stop sign/stop light detection. Nor will you see the car able to handle unexpected road situations (construction, obstructions, emergency vehicles, temporary detours, etc...) All this will likely be AP3 & AP4.

It's why I've decided to skip AP2 altogether. I won't add much that AP1 doesn't already do, except a few nifty parlor tricks. AP3 will be the big leap from AP1. I predict we'll see AP3 sometime in late 2018 or mid 2019.

 

[AnxietyRanger]

@Naonak: Indeed, AP1 is still missing many of its original promises. No arguments from me there. We may see some level of ramp-to-ramp with the final (?) major AP1 updates, but the level of summon or traffic light recognition promised I guess will never happen on AP1.

It does seem logical that AP2 might fall short on the full self-driving end of the promise-spectrum, namely the idea that your car could be out there driving without you in the car. I'm not saying the can't make it work in theory, the cameras are sufficient for that and the CPU/GPU can always be upgraded if need be... the problem is the redundancy and regulation needed for a non-driver car.

I just don't know how AP2 can manage in inclement weather or in many special scenarios with no 360 radar, no lidar, and no wipers for the side and rear cameras. Also the lack of a nose camera or nose corner radars does pose questions about many scenarios where the sides of the car are obstructed (e.g. close walls/tunnel) when maneuvering in tight quarters.

I don't have too hard a time seeing AP2 fall short on the full self-driving promise without a driver. Maybe they'll get it approved in some jurisdiction in nice weather, or while a driver is present as a backup, but as a money-making, ride-sharing autonomous car doing Uber type of rounds for you by itself on the Tesla Network...?

I can see that promise being broken and relegated to AP3 or AP4...

 

[electracity]

But the point is, MOST software ships on time because more and more resources get thrown at it if the timetable starts to slip.

Most software ships on time because the feature set is reduced. Like 8.1.

Musk apparently had to ship AP2 because he didn't have the right to continue to ship AP1. Considering AP2 status today it would have been better to ship AP1 longer. Later this year we will see more advanced autodrive from major manufacturers. But AP1 still looks good today, and Tesla could have avoided the AP2 kerfuffle.

 

[AnxietyRanger]

Regardless of legality, what was Musk's motivation for taking FSD money and suggesting the functionality was in the near future?

This situation is like the model X where Tesla held a fake launch. Tesla doesn't do fake product launches because Musk is optimistic. They do fake product launches because they feel they have to.

I assume they fake launched new autopilot because it was unacceptable to admit that they were taking a functional step backwards from AP1.

Fake is the new reality. A lot of press still reports Tesla as the autonomous driving leader. So I guess fake works. The model X turned out good, so maybe the new AP will be great.

The model 3 launch will be fake too, but that is another topic.

Launching Model X in late Q3/2015, even though they in reality could only start shipments in late Q4/2015 and even then in very limited fashion, allowed them to claim some bonuses, but I speculate most importantly bought them a full quarter of time from the market.

Same with AP2. With the announcement and delayed and very limited launch of AP2 software bought Tesla a full quarter. Even though it eventually shipped in a very limited fashion, the AP2 software, Tesla had already managed to weather the MobilEye storm at that point.

Both cases of weathering a stormy quarter by delaying bad news to the point where it no longer matters? Just speculation.

 

[Wolverinegeoff]

@Naonak: Indeed, AP1 is still missing many of its original promises. No arguments from me there. We may see some level of ramp-to-ramp with the final (?) major AP1 updates, but the level of summon or traffic light recognition promised I guess will never happen on AP1.

It does seem logical that AP2 might fall short on the full self-driving end of the promise-spectrum, namely the idea that your car could be out there driving without you in the car. I'm not saying the can't make it work in theory, the cameras are sufficient for that and the CPU/GPU can always be upgraded if need be... the problem is the redundancy and regulation needed for a non-driver car.

I just don't know how AP2 can manage in inclement weather or in many special scenarios with no 360 radar, no lidar, and no wipers for the side and rear cameras. Also the lack of a nose camera or nose corner radars does pose questions about many scenarios where the sides of the car are obstructed (e.g. close walls/tunnel) when maneuvering in tight quarters.

I don't have too hard a time seeing AP2 fall short on the full self-driving promise without a driver. Maybe they'll get it approved in some jurisdiction in nice weather, or while a driver is present as a backup, but as a money-making, ride-sharing autonomous car doing Uber type of rounds for you by itself on the Tesla Network...?

I can see that promise being broken and relegated to AP3 or AP4...

I think you are right about it being tough to meet the regulatory hurdles for driverless with current hardware. I wonder if they priced in a future hardware upgrade for whatever next gen Nvidia chip has more redundancy built in, or buying time for them to get their own chipset completed and ready for install.

As far as the weather, I can only speak to concerns related to rain, being a Floridian. If the camera can't see well enough, it's likely I can't either in the same situation and it may be smarter to just wait it out, or take over for the car. As for when it is driverless, I can imagine a scenario where they program it to pull over into a parking lot or side of highway until conditions improve. I don't fully buy in to the Tesla Network either, but I also don't think I'd want random people wearing down my interior or getting their grossness all over it. My guess is it won't have the ability to yell "NO EFFING SODA IN THE BACKSEAT!" at the passengers, so it's a no-go for me!

 

[AnxietyRanger]

I think you are right about it being tough to meet the regulatory hurdles for driverless with current hardware. I wonder if they priced in a future hardware upgrade for whatever next gen Nvidia chip has more redundancy built in, or buying time for them to get their own chipset completed and ready for install.

To me it seems possible Tesla may upgrade the FSD cars at some point with their own hardware, and may have factored this into the price. Though it is also possible they will try to cope with the existing hardware. But overall, it is not the CPU/GPU that I am worried about, I think that in itself will be sorted out eventually. As will the software, though that could take years... It is the rest of the hardware that troubles me more.

Well, not troubles exactly, as I am not really expecting too much of the suite. I am interested in following its progress, where ever it may lead.

As far as the weather, I can only speak to concerns related to rain, being a Floridian. If the camera can't see well enough, it's likely I can't either in the same situation and it may be smarter to just wait it out, or take over for the car. As for when it is driverless, I can imagine a scenario where they program it to pull over into a parking lot or side of highway until conditions improve.

I can see AP2 being approved for driverless driving in some progressive good weather states such as California - after lots of software maturity and perhaps the CPU/GPU upgrade mentioned above. Maybe this is the extent Tesla is aiming for in this generation anyway, to be able to do its thing on its home turf - by then AP3 or AP4 will be out anyway and tackle wider geographical areas...

As for concerns related to weather, I continue to be of the opinion that fixed cameras are much more susceptible to weather issues than humans are. First of all, humans can move their heads past dirt etc., and we can of course stop and clean the windows as need be. 5 out of 8 cameras in Tesla's suite only have heating to help them out, that will not sort many scenarios, which poor weather can make worse. And for wintery areas, snow can quickly wreack havoc anywhere, even with heating in place.

The lack of 360 radar and lidar also concerns me, as does the lack of a nose camera. These would offer more redundancy for many situations - the ultrasonics are too slow and short range to help in many situations.

Still, I expect the AP2 FSD will one day be a great driver aid... but the driverless part seems iffy for reasons listed.

 

[ecarfan]

I thought by now people would have reloaded not to take Elon's words at face value. He is largely wrong with both Tesla and SpaceX.

I don't know how to "reload" someone's words, but what Elon has accomplished with Tesla and SpaceX is clearly extraordinary: revolutionizing two long established and technically challenging industries in ways that many "experts" said was impossible.

It's not boasting if you do it. And he's doing it.

So @gowthamn, please share with us what you have done to decrease carbon emissions on a comparable scale, and also what you have done to improve the chances of humanity surviving the next existential thread to its survival?

I'm waiting...

 

[gowthamn]

I don't know how to "reload" someone's words, but what Elon has accomplished with Tesla and SpaceX is clearly extraordinary: revolutionizing two long established and technically challenging industries in ways that many "experts" said was impossible.

It's not boasting if you do it. And he's doing it.

So @gowthamn, please share with us what you have done to decrease carbon emissions on a comparable scale, and also what you have done to improve the chances of humanity surviving the next existential thread to its survival?

I'm waiting...

Don't use logical fallacies. You argument is stupid. What does me doing something to reduce carbon emission got to do with elon lying?
It has been proven again and again that Elon misses deadlines, lies and creates hype for both Tesla and SpaceX. Just look at his statements from a neutral perspective. Not from a fanboy perspective.
Yes, he has done what no other person had done before. He deserves credit for that. But he lies and creates hype. He should be criticized for that.

 

[Sir Guacamolaf]

I don't know how to "reload" someone's words, but what Elon has accomplished with Tesla and SpaceX is clearly extraordinary: revolutionizing two long established and technically challenging industries in ways that many "experts" said was impossible.

It's not boasting if you do it. And he's doing it.

So @gowthamn, please share with us what you have done to decrease carbon emissions on a comparable scale, and also what you have done to improve the chances of humanity surviving the next existential thread to its survival?

I'm waiting...

I thought we were discussing Elon's over-promising and misleading tweets - not his accomplishments.

But I don't mind you changing topics. I'd say Mother Teresa makes Elon look like a first grader.
Not only did she do a lot for humanity, she didn't make billions off of the backs of her followers or lie to or mislead them either.

So back to Elon's over-promising and misleading tweets, don't you think they cause a lot of negative energy?
I sure think they do. AP1 owners were pissed, and AP2 is the same movie all over again, just worse.

I know stock market has so far reacted positively, but only a matter of time before the chickens come to roost.

 

[ohmman]

But I don't mind you changing topics. I'd say Mother Teresa makes Elon look like a first grader.
Not only did she do a lot for humanity, she didn't make billions off of the backs of her followers or lie to or mislead them either.

Mother Teresa was a masochist who glorified the suffering of others. If you're a Catholic, you may not agree, but the rest of the world would say she certainly misled her followers. She led them into her version of heaven, through the path of abject poverty and suffering.

Doing something to stave off climate change - a worldwide phenomenon that data suggests will cause tremendous suffering - fits more broadly into an agnostic version of doing something for humanity. Whether one gets rich off of it or not is beside the point.

 

[KrenGrl]

And you as a software engineer have never given an estimate to a boss that turned out to be too short?

You've never had a task end up being more complicated than you thought it would be?

Answer honestly.

I regularly contract for and receive said estimates from software engineers. They are never on time, or under budget.

You clearly know Elon's time warp. Why are you angry if you know the history? Fool me once I think?

I'm also convinced that if everybody who worked for him was as smart as him or worked as hard as him, his time lines would always be accurate.

Lighten up.
-Jim

Same experience here. The only time that my estimates are wrong is when I go shopping or a software engineer gives me an estimate.
Otherwise, I plan well and don't get behind.

 

[AnxietyRanger]

If you're a Catholic, you may not agree, but the rest of the world would say she certainly misled her followers.

Now replace Catholic with a suitable word and she with he and you are certainly describing Elon.

 

[Sir Guacamolaf]

^^ This thread has clearly gone off topic. LOL

 

[AmpedRealtor]

Regardless of legality, what was Musk's motivation for taking FSD money and suggesting the functionality was in the near future?

I suspect it's not much different than accepting deposits for unreleased vehicles, like the Model S, Model X, and Model 3. It's also not much different than taking your money for next-gen seats but then delivering a car with last-gen seats plus a "due bill" for seats to be installed later. It's kind of like what Tesla did with AP1, selling the sizzle but not delivering the steak for almost a year after taking orders.

What you describe is normal policy at Tesla. The reason is quite obvious: Generate revenue and use that revenue to develop the promise. I wouldn't doubt if EAP (let alone FSD) did not exist at the time Tesla took deposits and was going to use a large portion of that deposit money to develop the promised features. Breakup with Mobileye? Give me a break! What did Tesla expect Mobileye to do when it proposed stealing Mobileye's intellectual property? Tesla wanted to run AP1 in parallel with AP2 so AP2 could "learn" behavior from AP1 - in other words, reverse engineer Mobileye's technology. Good on Mobileye for saying NO to outright theft, and they are in a much better position today to deliver on the promise than Tesla.