Model S

Full Self Driving - what are we getting?

edited November -1 in Model S
When ordering a Model S or X, Tesla appears to be promising that purchasing Full Self Driving will enable "full self-driving in almost all circumstances", with the caveat that the functionality is dependent upon extensive software validation and regulatory approval.

Until Tesla has completed the validation, it's possible Tesla will determine that the AP 2.0 hardware/software may never deliver FSD or that it may never receive regulatory approval.

And yet, Tesla is asking owners to pay $3,000 up front (or $4,000 later) for functionality that may never be delivered.

We're planning to order a 100D with AP 2.0, and will likely check the box for FSD. However, I wish Tesla was being more realistic about the FSD option.

A few recommended changes for this option:

For $3,000, Tesla commits to provide FSD. If additional hardware is required to complete validation or get regulatory approval (such as the addition of LIDAR or the ability to communicate with other FSD vehicles), Tesla will add that hardware for no additional charge.

If FSD is not activated at purchase, it MAY be available in the future for an upgrade fee. If no additional hardware is required to achieve FSD, the upgrade will be $4,000. If Tesla has to install additional hardware to achieve FSD, then the upgrade is $4,000 plus the cost of the hardware upgrade(s).

Before FSD is enabled, Tesla will extend the Enhanced Autopilot functionality to provide driver monitored autopilot on urban streets, in addition to EAP's support for highway driving.

At this point, Tesla believes they have sufficient hardware to get FSD working - though until they've actually done the testing, they can't be sure they won't need additional hardware to make it completely work. Plus, they don't know what additional hardware may be needed to achieve regulatory approval (and addressing liability).

The above strategy would be more realistic, based on where we are with FSD technology and regulation today. It would provide owners purchasing FSD today assurance Tesla will guarantee they will (eventually) get FSD, even if more hardware is needed. And for those who want to wait, they'll face the potential of additional hardware costs if they delay their FSD commitment.

Since Tesla appears confident they've got what's needed for FSD with AP 2.0, then providing a more detailed commitment on FSD shouldn't be a change in direction - just a clarification of what they are committing to people willing to spend $3,000 today on future functionality.


  • edited November 2016
    The only "upgrades" included after purchase are software updates.
  • edited November 2016
    Great question. Tesla has already determined that some of the initial promises for 1.0 are beyond the hardware's capability, so have pulled back in some areas and limited others, will never deliver some. There appears to be a bit of the same guessing with 2.0, so I am on upgrade hold until the same functionality appears that we already have with 1.0 hardware and 8.0 software. No point in retrograding just to wait for TACC, AEB, etc. For the first time, Tesla has riddled its announcement with caveats that basically allow it to deliver whatever, whenever, with no liability for what it learns in development. Caveat emptor.
  • edited November -1
    Best to follow this rule: Don't pay for something you aren't getting. Follow that rule and you won't be disappointed, and threads like this become unnecessary. I certainly wouldn't pay for a feature that was "coming in the future". Given Tesla's track record of broken promises to Autopilot 1.0 owners, I have little confidence that Tesla can deliver what it's promising. I hope they can, but based on their track record of overselling Autopilot 1.0 features that never came true, it's not looking good.
  • edited November 2016
    +1 ^
  • edited November -1
    Tesla needs to deliver full self driving capability unlocked in each jurisdiction, as it becomes legal.

    That is what they are promising with current hardware.

    If that isn't permitted by law, Tesla has "delivered" as promised.

    If the vehicle isn't capable, but is permitted, Tesla better be prepared to write some refund checks.

    I'm believe Tesla has produced ALREADY produced the first passenger vehicle, with enough mileage to validate hardware, permitted to operate in self driving / autonomous mode.
  • edited November -1
    Another thing to consider when upgrading later if you so choose is the possible savings in sales tax. In my case, NY, if the $3k option is included in the sales price when I register the car, that equates to an additional $240. So I effectively would be paying $4k-$240=$3,760 if I upgraded after it's proven worthwhile. This is assuming NYS doesn't find a way to tax me at time of purchasing the upgrade post registration.
  • edited November 2016
    I paid for future software delivery in my 12/2014 P85D, but the software had been fully demonstrated in the roll-out, and I was delirious on pain meds in the hospital when I placed the order. The functions were clearly laid out and mostly feasible, so I decided to move ahead to get 4wd, etc. Although I said before the car was ever delivered that some of the promised functionality was impossible with the 1.0 hardware. I also said that I did not care, hoped they could prove me wrong and actually develop functions like hands-free exit to exit, summon to curb, etc. I was therefore not disappointed when some of the promises went away or were retracted - it is what I expected all along.

    On the other hand, this version is far more speculative regarding whether it will ever work, and how. My driveway is a mile long gravel with undefined edges and several dogleg turns, deep ditches nearby. Good luck with fully automating that. I hope they can. will buy it when/if ready, but am skeptical now. I think that Tesla THINKS it can do this, is unsure, will figure it out as it goes along. In reality, it has some idea how the new hardware works, is probably a long way from refinement. Again, I do hope to be wrong on this, but it is Tesla's history, and they are further out over their skis on this than with any prior commitment. Previously they were just stringing together Christmas lights with some linking/processing software, getting others with a much larger volume mandate to develop the sensor firmware and stuff like reading signs. Now it is all on Tesla, and one must question whether their sales volumes, even at 1 million per year, can possibly justify the gargantuan task of putting out an artificial intelligence machine, which is what they say is now intended. It shall be interesting, unless the SCTY fiasco takes the company down before the holy grail is reached.
  • edited November -1
    If they are correct, Tesla will be the most valuable carmaker EVER, perhaps the most valuable company. So it is worth the risk to take the self driving market, fast.

    Perhaps they are like Icarus, too fast and hot. Or they actually do something HARD.

    I hope it's the latter.
  • edited November -1
    So @PD, when are you ordering? '-)
  • edited November 2016
    Maybe I'm reading it wrong, but EAP is $5,000 and that is subject to the same disclaimer. Thus, if you just purchased EAP, you are not getting the same functionality that exists today, but will get BETTER functionality if/when approved. Further, you can spend $3,000 (for a total of $8,000) for FSD, also subject to approval. So it is conceivable that you spend $8k, not just the $3k, and never have those features? Has there been any commitment that if you spend the $5k for EAP and the additional functionality is NOT approved, you would at least have the same basic functionality you get today with EAP? Just doesn't seem right ...
  • edited November 2016
    FSD and P100D, sorry, pretty sure if FSD will ever come, in FSD mode the S60 and P100D will drive the same, they will not race! I think the FSD quality will depend much more on roads, markings, signs quality rather than software quality. In any case FSD will drive like grandpa.
  • edited November 2016
    Tesla hasn't stated that they won't deliver a feature that lets it meet me at the curb in front of my house. The way it works now, it seems to undo whatever it did on the way into my garage when it backs out. It seems quite feasible for them to get it to learn a specific path, such as the one from the garage to the curb in front of my house. I don't really see a need for it though, but they did promise it.

    As for hands free entrance to exits on freeways, it's not that they can't do it. They added the nag even though it's not necessary for the car to perform. They haven't shown evidence that drivers are more attentive when they have a hand on the wheel at all times. They could change this at a later date if data shows that it doesn't prevent accidents. It's already clear to me that drivers get into accidents all the time due to inattentiveness in other cars, and they almost always have their hands on the wheel.
  • edited November 2016
    Don't confuse two things:
    With the EAP package you get the 8 camera set and - by late December - a very much enhanced autopilot. With the added 4 side cameras Autopilot will be even better.
    Only the license to run autonomously is the unknown element here.
    That is the way I read it...
  • edited November 2016
    And as far as the hands on Nag is concerned, Haggy, the question is not whether Tesla can prove anything but a question of driver behavior - and ultimately legislation.
  • bpbp
    edited November -1
    According to the Tesla website, EAP increases the number of active cameras from 1 (AP 1.0) to 4, plus 12 ultrasonic sonar sensors and the new NVidia processor. EAP provides driver assistance for highway driving - lane keeping, automatic lane changing, freeway transitions, exit ramp, self-parking and summons to/from a garage. These are driver assistance features, requiring driver monitoring - and they expect this to be operational next month.

    AP 1.0 and EAP are supposed to work only on highways - and for safety reasons, Tesla could impose geofencing restrictions so those features only work on highways - and when the cars are on surface streets, the features are disabled.

    FSD activates the other 4 cameras, and is supposed to work in almost all driving conditions - not just high speed driving on highways, along with enhanced parking and summons. While the description talks about FULL SELF DRIVING, what isn't mentioned is that Tesla is likely to enable these features in a "driver assist" mode before the software has been fully validated and has regulatory approval.

    If Tesla starts enabling the FSD features - in driver assist mode, that does provide increased value over EAP, even if Tesla never gets approval to enable full self driving.

    Tesla could and should provide more details on their plans for this feature, since they are asking owners to make a purchase commitment for a feature that they haven't developed yet - and may never work technically (it may need more sensors) or may never receive approval.
  • edited November -1
    If I was ordering today I'd probably go for the enhanced autopilot since in theory it will be available "soon". I'd wait on full self driving because god know how long the government will stall that .
  • irir
    edited November 2016
    Considering the havoc that a drop of water wreaks on the backup camera, and after it evaporates the dirt causes bad ghosting at night. I am concerned that more cameras, instead of more radar are the enabling factors for self-driving.

    Looking forward to AP 3.0 when the early adopters help work through the kinks.
  • edited November 2016
    @mathwhiz "So @PD, when are you ordering? '-)"

    It's up to Tesla. I will place the upgrade order the second that Tesla deliver at least TACC, Lane Keeping, Lane Change, and AEB to the 2.0 AP cars. My order will be for whatever is the largest battery and power combination at the time, and will be placed within seconds of being made available. Until then, it is vapor.
  • edited November 2016
    There's something to be said for taking the $3000 for full self-driving, buying Tesla stock, and assuming that by the time it's released in your jurisdiction, the stock will be worth well over $4000 even after taxes on the gains. Given that many people don't believe it will ever really be fully autonomous, delivering on that promise could give a good boost to the stock price by itself.
  • edited November 2016
    As @PD is familiar with contract law, he is no doubt familiar with the term "substantial performance." In my opinion, AP 1.0 has met that bar, though like others I hope for more complete summon functionality and maybe recognition of stop signs and red traffic signals.

    @PD, you've softened your early rather forceful prognostications about AP being dangerous and beyond the hardware's capability. I appreciate your partial but welcome turnaround that AP 1.0 isn't a resurrection of Stephen King's "Christine." :-)
  • edited November -1
    okay rxlawdude, uncle

    I never said that 1.0 actual vs 1.0 promised is sufficiently tortuous to generate lawsuits, although the Norway settlement is an interesting example of Elon stepping on his...
  • edited November 2016
    This reminds me of television technology. I spent a lot of money because I wanted to be the first person on the block to have HDTV--- not realizing how long I would have to wait to get a signal in HD. It took about 3 years. When it finally arrived, I gave away my TV because it was obsolete and bought a new one.

    Any take home lessons here?
  • bpbp
    edited November 2016
    A major difference between the HDTV (and later 4K TV) introduction and FSD is that the standards for HDTV/4K TV were (mostly) defined and it was only a matter of program providers and electronics manufacturers to fully implement those standards and provide the programming that could take advantage of the new technology.

    EAP is AP 1.0 using the new sensors, so that technology is likely to be operational (and soon).

    But with FSD, the technology is still under development, and until vehicles have been validated to successfully self drive, FSD with the AP 2.0 hardware is not guaranteed to ever work, and may require additional hardware to overcome as yet undiscovered limits that prevent FSD from working. And then there is the regulatory hurdle, which could impose additional requirements - such as requiring that all FSD cars communicate with each other. And, after that, there may be an additional hurdle to address liability concerns, since Tesla will likely push hard not to accept liability for FSD accidents.

    When I order my P85 4 years ago, it was like buying an HDTV. When my car was delivered, the nearest Service Center was 1100 miles away (UPS delivered the car to my front door) and there wasn't a supercharger network. But Tesla was early in delivering Model S cars - and they had plans in place to fully support the technology.

    FSD is different - it's more speculative, and Tesla isn't providing any guarantee it will ever work, even though they are charging $3,000 to add that to new cars.
  • edited November 2016
    <em><strong>There's something to be said for taking the $3000 for full self-driving, buying Tesla stock, and assuming that by the time it's released in your jurisdiction, the stock will be worth well over $4000 even after taxes on the gains. </strong></em>

    Remember that capital gains taxes are now 20%. Plus, if like many Tesla owners you are "rich" by the gummint definition, there's an additional 3.8% on top of that to fund the ACA.

    So your $3000 investment Tesla stock would have to be worth $4952 when you sold to leave $4000 after taxes.
  • edited November -1
    We can get Trump to do something about those nasty capital gains taxes. Shouldn't taxes be on earned income only? How could any decent Republican argue with that? If he asks why I want the change, it's because I'm smart. Or maybe selfish. It depends on whether I'm running for office.
Sign In or Register to comment.