Wait on adding AP2 and FSD or you will be sorry...maybe

[Drone Flyer]

I would expect Tesla to greatly increase the price of the FSD package once it starts offering meaningful capability. Sure those who bought in early are kickstarting Elon's self-driving system, but I think there's a bit of an early adopter discount baked in

FSD will never ever happen folks. Get over it!
Jurisdictions won't allow it for 20 years!
If you think AP 2.0 will be a complete self driving program, you are misinformed.
I really feel sorry for those who traded in their AP 1 cars for an AP 2 one.

 

[azred]

Sure, if your time is valueless to you and you want to spend several months driving manually - says I as I type this while AP2 pilots me down the road at the end of a 500 mile day that left me relaxed and feeling like I haven't driven even five miles.

My time is much too valuable to drive 500 miles, though if I did I would have too much common sense to post messages to this board while driving.

 

[chillaban]

FSD will never ever happen folks. Get over it!
Jurisdictions won't allow it for 20 years!
If you think AP 2.0 will be a complete self driving program, you are misinformed.
I really feel sorry for those who traded in their AP 1 cars for an AP 2 one.

Hey I traded in mine knowing and writing one of the most comprehensive reviews of the difference, and I don't need anyone to feel sorry

But in all seriousness: I do agree that Elon's vision of hands-free SF to NYC is not in the realistic future. But with that said, the FSD package could very likely hold some features people really want, such as lower nag intervals, stop sign and traffic light aware TACC, automatically taking turns in the city, etc etc etc, marketed as a really advanced ADAS beyond EAP's advertised features.


But all of this is just speculation. Unless one of the guys above is a Tesla employee trying to feed us covert information, I feel like we're wasting a lot of time speculating on the future when we've got a lot of work cut out right now understanding the features and quirks we have today!

 

[xyeahtony]

go ahead and buy EAP today. I use autosteer at 80mph on my commute to work everyday with no issues. (30 miles one way). Well worth it. I also used it on my delivery day trip from Houston to Baton Rouge.

 

[suraj1194]

But in all seriousness: I do agree that Elon's vision of hands-free SF to NYC is not in the realistic future. But with that said, the FSD package could very likely hold some features people really want, such as lower nag intervals, stop sign and traffic light aware TACC, automatically taking turns in the city, etc etc etc, marketed as a really advanced ADAS beyond EAP's advertised features.

But that would be a massive comedown, wouldn't it ? Take traffic light aware TACC. Mobileye documentation states that this functionality is capable from EyeQ3, i.e., AP1 . Sure, it's not there in AP1 yet, but the underlying system is capable of it. It's availability as an AP2+ feature in the form of one of the FSD capabilities would be a poor value proposition, amounting to artificially adding a value tier that could have been implemented within prior generation hardware. It's really not worth $3K on top of EAP if it just adds functionality thats technically present in the $2.5K AP1.

FSD should ideally involve use of all 8 cameras such that the car is capable of traffic ingress/egress awareness in 360 degrees. This means everything from autonomous lane changes in freeways, to being able to see and yield to bicyclists in the bike lane in city streets while turning.

 

[chillaban]

But that would be a massive comedown, wouldn't it ? Take traffic light aware TACC. Mobileye documentation states that this functionality is capable from EyeQ3, i.e., AP1 . Sure, it's not there in AP1 yet, but the underlying system is capable of it. It's availability as an AP2+ feature in the form of one of the FSD capabilities would be a poor value proposition, amounting to artificially adding a value tier that could have been implemented within prior generation hardware. It's really not worth $3K on top of EAP if it just adds functionality thats technically present in the $2.5K AP1.

FSD should ideally involve use of all 8 cameras such that the car is capable of traffic ingress/egress awareness in 360 degrees. This means everything from autonomous lane changes in freeways, to being able to see and yield to bicyclists in the bike lane in city streets while turning.

AP1 is already a massive comedown from what Elon originally announced. Do we really believe FSD will be different?

I was mainly entertaining the hypothetical that laws about attentive drivers and hands on the wheel will prohibit FSD from happening.

 

[suraj1194]

AP1 is already a massive comedown from what Elon originally announced. Do we really believe FSD will be different?

I was mainly entertaining the hypothetical that laws about attentive drivers and hands on the wheel will prohibit FSD from happening.

I keep reading about AP1 promising exit-to-exit control , but has there been any published reference on the Tesla site ? It's hard to parse Elon because he has no sense of time or continuity, and you can't tell if he's talking of an upcoming feature or his vision of the future.

I agree with you that regulatory hurdles will ultimately impact FSD usage the most. But from a value for money perspective, it needs to place a lot on the table to be worth the added premium.

 

[oktane]

I bought EAP & FSD when I got my car in December. I must say I'm disappointed at the pace of the roll out. Some may say that they have made great progress. And that is true now that we know they started from scratch after MobilEye. However, from my expectations in December, based on what we were told at the time, it has been disappointing.

I look at the video Tesla released for self-drive (on the streets to the office) and I wonder why is I cannot get anything close to that? I look at other car vidoes (Bolt for example Watch GM’s self-driving Chevy Bolt EV prototype slow down for a raccoon [Video]) and ask myself why can I not get anything close to that?

What is the answer? Are these videos so staged that they they are not even close to current capability?

It seems like Tesla puts out releases so often that we are getting their latest and greatest. If that is true then their automation video was fully staged. I don't know how else to interpret that.

All the others seems to take a totally different approach. They want their driving automation to be fully baked before release.

Yes, the FSD video was totally phony. It was hundreds of trials stitched to make one cut. Witnesses corroborate this, as do the CA DMV records from the autonomous trials the three days prior. All data is public and available. In total, Tesla had tested only a couple hundred mikes before releasing the video. FAKE NEWS to steal from customers!

I invite Tesla to load the same firmware they used on any production car and sit a reporter in the backseat and see what the car really does. Won't happen anytime soon.

 

[bob_p]

When developing new software, it's not unusual to develop a limited functionality prototype as a "proof of concept". It can look close to the real thing, but have severe limitations, often able to provide only enough functionality to perform the "proof of concept" demonstration.

The FSD video feels like a "proof of concept" prototype - something that shows the hardware sensors and processors are capable enough to support FSD, but with software that was only able to handle a specific route - and under certain conditions. Tesla's claims that the AP2 hardware is sufficient to support FSD may have been based on testing that included the FSD demo.

Like with other POC prototypes, for those not involved in the project it's easy to assume that since the prototype appears to be operational the software is close to ready - when in reality, the development team knows they took many shortcuts to get the demo working - and almost all of the software development is still in front of them before they have an operational system.

The FSD video does give the false impression that Tesla was close to having FSD working, and they could have provided a better explanation of what the demo is showing, and the limitations in the software (note that on the display, the navigation software is not showing a route for the demo run).

Since some customers are disappointed in their purchase of EAP and FSD, Tesla should offer those customers refunds of their EAP and FSD options, features that Tesla has not yet delivered. Though, if they accept the refunds, if/when EAP and FSD are operational, they'll have to pay whatever the prevailing upgrade price may be - which will likely cost them more in the future.

 

[oktane]

Exactly, but they refuse to even acknowledge this is even an issue. From the highest levels, they insist owners "be patient" without being transparent on timeline and probabilities. Hence the anger. Patience has its limit.

 

[Viking1]

The FSD video feels like a "proof of concept" prototype - something that shows the hardware sensors and processors are capable enough to support FSD, but with software that was only able to handle a specific route - and under certain conditions. Tesla's claims that the AP2 hardware is sufficient to support FSD may have been based on testing that included the FSD demo.

In other words, it was like the facade of a building in a Hollywood studio lot. It looks like a real street, but once you look behind the facade you see there is nothing there. I guess "proof of concept" is one phrase you can use. I can think of others.

 

[stopcrazypp]

Yes, the FSD video was totally phony. It was hundreds of trials stitched to make one cut. Witnesses corroborate this, as do the CA DMV records from the autonomous trials the three days prior. All data is public and available. In total, Tesla had tested only a couple hundred mikes before releasing the video. FAKE NEWS to steal from customers!

I invite Tesla to load the same firmware they used on any production car and sit a reporter in the backseat and see what the car really does. Won't happen anytime soon.

This is incorrect if you are referring to both videos. The first video was cut, but a vast majority of the failed attempts were related to weather (it was raining most of the days). If you look at the video they didn't use any footage when it was raining, so they were mainly cutting and using footage from the last dry days in the October attempts. There were only 13 disengagements during the dry days.
Autonomous Vehicle Disengagement Reports 2016

The last video they released however was uncut and there were no failed takes (the November report had 20 miles but no disengagements).

 

[stopcrazypp]

I bought EAP & FSD when I got my car in December. I must say I'm disappointed at the pace of the roll out. Some may say that they have made great progress. And that is true now that we know they started from scratch after MobilEye. However, from my expectations in December, based on what we were told at the time, it has been disappointing.

I look at the video Tesla released for self-drive (on the streets to the office) and I wonder why is I cannot get anything close to that? I look at other car vidoes (Bolt for example Watch GM’s self-driving Chevy Bolt EV prototype slow down for a raccoon [Video]) and ask myself why can I not get anything close to that?

What is the answer? Are these videos so staged that they they are not even close to current capability?

It seems like Tesla puts out releases so often that we are getting their latest and greatest. If that is true then their automation video was fully staged. I don't know how else to interpret that.

All the others seems to take a totally different approach. They want their driving automation to be fully baked before release.

Your characterizations does not match how Tesla does updates. Tesla's releases are not directly the "latest and greatest". There is an extensive amount of in house testing before it is released in the public. For example, from the earnings call in February Elon mentioned he was already driving high speed AP2 3 months ago (around November last year). However, high speed AP2 did not get released until end of March.

And that is for something relatively simple. Something as complex as FSD, I would imagine needs a lot more in house testing before it is going to be ready and thus more lead time.

 

[stefan.bc]

There is an extensive amount of in house testing before it is released in the public.

If true most False Collision Warnings on most major highways would have been whitelisted before AP2 was released. I had to many to feel safe using the TACC.

 

[oktane]

@stopcrazypp Which is the second video? Thx

 

[stopcrazypp]

@stopcrazypp Which is the second video? Thx

This is the second video taken in November (according to report there were no disengagements when they made this video):
Tesla Self-Driving Demonstration

This is the first video from October which had cuts:
Full Self-Driving Hardware on All Teslas

 

[stopcrazypp]

If true most False Collision Warnings on most major highways would have been whitelisted before AP2 was released. I had to many to feel safe using the TACC.

False positive results can exist no matter how much testing you do. I don't think a white list works for this (it only works for overhead signs, but not in general). If you look at forums about other warning systems, they also have plenty of false positives, to the point the owners just turn them off.

An example here from Honda I read about recently in an article about their lane departure warning:
Lane departure warning gets turned off

 

[stefan.bc]

@stopcrazypp I was talking about overpasses, metal structures, radars, cameras on major highways where legal speeds exceed 60mph and using TACC. I was only replying to your comment that Tesla does extensive testing. It doesn't look like they did.

And on top of that there was no way to disable the TA part and go back to the regular CC.

 

[bhzmark]

My time is much too valuable to drive 500 miles, though if I did I would have too much common sense to post messages to this board while driving.

Doing calls in the car or listening to podcasts is a better use of time than standing in a tsa line or being cramped on a regional jet doing short flights.

When on the highway, and especially when following a lead car, AP works great to allow email and other phone tending especially when phone dictation works well through the Bluetooth car microphone. But still have to watch for those partial half in the lane blockages that a lead car might swerve around and similar situations.

 

[boaterva]

It seems like every one of these threads ends up with or steers close to (no pun intended) 'FSD will never happen'. But, as is stated on the website, FSD is supposed to have features way before Level 5. One other recent mega-thread was about Elon 'lying' because FSD hadn't 'deviated' from EAP 'yet' when his last prediction (for what it was worth) said FSD-specific features would appear between late April and late July this year.

No matter the scheduling and actual appearance of the features, we have to remember that FSD is not the 'robot car, I can sleep in the back', it's also the 'other features' above EAP all the way up to Level 5. I do wish just *one* would appear so we could move on, just a little!