TMC is an independent, primarily volunteer organization that relies on ad revenue to cover its operating costs. Please consider whitelisting TMC on your ad blocker and becoming a Supporting Member. For more info: Support TMC

2 years: Musk's new timeline for door-to-door full autonomy

Discussion in 'Model S' started by calisnow, Apr 28, 2017.

  1. phaduman

    phaduman Supporting Member

    Joined:
    Sep 16, 2016
    Messages:
    368
    Location:
    San Jose
    Tesla should live stream the LA to NYC trip of the self-driven demo car - just like the live telecast of the free fall from space show not so long ago.
    They should announce the date before hand, build up excitement...fanboy dream come true and also for the "i told you so" crowd - who knows, the president might watch that whole week-long streaming.
     
    • Like x 2
    • Love x 1
  2. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    Considering that several companies have already driven from coast to coast in autonomous cars, I don't think anyone doubts Tesla will be able to do it. Only two miles of it are city streets anyway.

    We will certainly see that demo at some stage. Possibly sooner than the Roadster 3.0 roadtrip demo. :) And maybe even around end of this year.

    After all, as Elon Musk says, Summon from NY to LA should be possible around January 2018. :)
     
    • Funny x 3
  3. Zetopan

    Zetopan Member

    Joined:
    Oct 10, 2015
    Messages:
    100
    Location:
    Hillsboro, Oregon
    AHA!! Just As We Thought! The AP2 car will have to travel So Slowly that it will take TWO YEARS to reach your destination! [/AngstRiddenPosterEmulationMode] ;)
     
  4. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    No no, haven't you read Elon, you Summon the car from LA to NY, whichever you are, in ~January 2018. You can then drive yourself to your destination. Problem solved.
     
  5. essmd

    essmd Phantom X

    Joined:
    Aug 14, 2015
    Messages:
    372
    Location:
    DFW, Texas

    I can see you are a Tesla fan... so am I. In fact I currently plan to only drive a Tesla in my next 30+ years of driving, the latter of that probably in a self-driven one since I know that even my AP1 adds so much safety when activated, I expect FSD to be even more so.

    The promise of OTA updates is unique and amazing, and in the beginning there was a giddiness when receiving an update, hoping for improvements, new features and conveniences, etc. But for my HW1, there now seems to be feeling of abandonment, as if Tesla is indicating that my car that was delivered in July, is now as good as it is going to get, so it's now like any other manufacturer's vehicle in that respect. I hope I am wrong, and probably spoiled by the first 6 months, but perhaps not.

    I do know that the day HW2 was announced, I just lost about a 1/3 of my car's value, since selling a HW1 will be a real challenge n the future given the newer hardware, but that can happen with any vehicle, just usually with the next model year.

    Ironically, I also was given an opportunity to sell my MX back to Tesla and walk away based on an unfixable abnormality, yet I chose to order another one, so my AP2.0 is about to be built. This gives me an opportunity to get the AP2, keep the unlimited charging, even save some money, but mostly I can actually have the only real SUV Tesla sells, the 5-seat config, which restores the original promise of folding 2nd row seats with a seat back that actually can be extended for passenger comfort, not just tilted back.

    But I am not blind to the flaws of Tesla, both physically with the MX (fit and finish is subpar to premium vehicles costing half the price), as well as having insight into the dysfunctional handling of some customer satisfaction issues.

    However, now there is a new issue that I'd love to get your perspective on. It has to do with the elimination of the 2nd HVAC fan in the 5-seater, which combined with rerouting of the ducts and vents results in a severely anemic airflow to the 2nd row passengers, such that getting into a hot cabin and cranking the A/C will not be nearly effective as the other configs, or probably any other SUV on the market,

    The thread is

    How adequate is the A/C airflow in the MX 5-seater for 2nd row passengers?

    You give Elon a pass for doing what he has done so far, I will hold that judgement until he and his leadership resolve this issue, as it was a huge mistake to make this design change, and they are guilty of not considering owners that live in hot summer climates, beyond the borders of California.

    Any guess on how many 5-seaters are living in parts of the country with hot summers??? Yup, its many.

    Let's see how your boy handles this one. I don't judge a man on how well he does something right, but rather how he corrects something he did wrong.
     
  6. lunitiks

    lunitiks Cool James & Black Teacher

    Joined:
    Nov 19, 2016
    Messages:
    2,708
    Location:
    Prawn Island, VC
    Elon's stated many times that if the PX2 isn't up to the task -- it's "easy" to upgrade it. Why is the CEO so outspoken about this? (Talk about massive upgrade costs! Tens, if not houndreds of thousands M3s will be on the road with this HW by the time FSD is finalized.)

    Why isn't autodimming and auto wiping ready yet? What's going on?

    I'm starting to suspect that AP2 is a (very high stakes) experiment. Hope I'm wrong
     
    • Helpful x 1
  7. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    #67 AnxietyRanger, May 1, 2017
    Last edited: May 1, 2017
    The CPU upgrade is the least of my worries, though. I am expecting it to be factored into the FSD price anyway. Sort of like with EAP you are paying for the PX2 plus premium. With FSD you are paying for a future board of perhaps Tesla's own silicone (as speculated, if upgrade is needed) plus premium...

    Personally I worry more about how FSD can they get (especially in inclement weather) without the redundancy of 360 radars/LIDAR. They only have redundancy (multi-cameras and radar) basically on the front as well as elsewhere at very low-speed/close-up scenarios through the ultrasonics. We shall see.
     
  8. essmd

    essmd Phantom X

    Joined:
    Aug 14, 2015
    Messages:
    372
    Location:
    DFW, Texas
    There is a huge distance between Tesla saying "here it is, our autonomous driving vehicle, just watch the video of the passenger sleeping form door to door" AND the regulators approving a Tesla to do just that.

    And what happens to all of the owners that plopped down the $3K, but for whatever reason (lack of LIDAR?, Need for HW3?), municipality approval never occurs???

    Will Tesla simply refund that $3K, or will they shrug shoulders and say "Hey, we built it, it works, not our fault it wasn't approved, we even mentioned that possibility in the disclaimer you read before you submitted your order."

    Just watch all the Fanboys jump on and say "You can't blame Tesla for governmental bureaucracy"

    At least you can summons your Tesla to meet you at on the driveway somewhere...
     
  9. essmd

    essmd Phantom X

    Joined:
    Aug 14, 2015
    Messages:
    372
    Location:
    DFW, Texas
    #69 essmd, May 1, 2017
    Last edited: May 1, 2017
    The weather is one thing, but I know my AP1 can see the lines in a dark and stormy night much better than i can. it has to be better with the newer camera design.

    I think a bigger concern is those parts of the road design that will challenge any autonomous driving, constantly changing road construction (lane shifts, orange barrels encroaching the lane, when another car hits a FSD Tesla, or when the car hits something... car, structure, person???

    I feel that as they get closer to saying "Almost Done", there will be more challenges identified that they did not account for, like for every 1 steps forward, it's 2 steps back.

    And the LIDAR issue will be there always, comparing the functionality of vehicles with to those without, or perhaps some other new technology.
     
  10. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    IMO with FSD Tesla can probably get away with that. If it is really regulatory that is stopping them and not failure to get the car otherwise FSD-like functional.

    I am more concerned about how EAP was marketed to us. That is the story where Elon/Tesla once again were quite reckless with their messaging.

    Of course if it turns out AP2 can't ever do FSD even with a driver there as a backup, that would be a problem for Tesla. And they have made some public statements on the timeline on when that is expected to start happening, so on that they indeed need to deliver...
     
    • Like x 1
  11. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    #71 AnxietyRanger, May 1, 2017
    Last edited: May 1, 2017
    I don't worry about dark so much as dirt and snow. The only remedy against that they have is camera heating around the sides and back. A driver can move their head to see around dirt, but with a small, stationary camera, dirt seems like a much bigger potential issue.

    Not just LIDAR, but also the competition having 360 radar. Tesla has no redundancy on the sides or around the back (excluding low-speed, close-up).

    I am speculating AP3 will add that side and back redundancy. Though of course a small chance remains Tesla is religiously committed at going visual, even to the extent that they might remove the radar from Model 3. But I'm still thinking mostly AP3 will add more radar at the very least, perhaps more cameras and possibly (but IMO least certain) LIDAR.
     
  12. essmd

    essmd Phantom X

    Joined:
    Aug 14, 2015
    Messages:
    372
    Location:
    DFW, Texas
    Ah, living in Texas I forgot about the snow and ice, but for a few days every winter, It can be an issue.

    You are referring to dirt and snow on the windshield impairing vision, as I am referring to it on the road, but either perspective, that can be a real challenge.

    Did not know about 360 radar, that is that newer technology I did not know about, so what else is out there?

    One thing not recently mentioned is "fleet knowledge", or I call it Hive Intelligence, where learned elements are uploaded from individual vehicles than transmitted to all that hit approach that GPS location. Is that still a thing?

    There has to be some process where a car enroute on a trip can no longer continue for whatever reason, How will Tesla reconcile with that?
     
    • Like x 1
  13. AnxietyRanger

    AnxietyRanger Well-Known Member

    Joined:
    Aug 22, 2014
    Messages:
    9,451
    Location:
    EU
    Sure, both perspectives can be challenging, but my main point was the concern about dirt (or snow) impairing camera vision. Not just on the windshield either (especially as the Tesla front has wipers and redundant cameras and a radar), but especially on the sides and back of the car. I am more hopeful that dirt on the ground will eventually be sorted out by the neural networks. But they do need to see...

    Not necessarily new technology, just something that other manufacturers have been doing and Tesla has not. Already by 2010 Audi A8 had four radars, in each corner of the car. I guess it wasn't 360 degrees quite yet, but it was many degrees anyway. Mercedes Benz has been shipping cars with I think six radars. Just to run their driver's aids. A lot of the European efforts around self-driving platforms that have been publicizes are based on 360 radar, 360 vision, 360 ultrasonic plus LIDAR in addition (e.g. Volvo as well).

    I guess it is, certainly it can add a layer of data to the mix.

    That's a good question. A constant "plan B" for parking the car safely might be a part of the car's thinking constantly. While better than disengaging or crashing, that still won't be very helpful if the car was driving itself to pick you up. That said, I can see a new kind of road service developing around such instances, as long as most of the time the car works fine by itself...

    I just worry if AP2 has the optimal mix of sensors yet. I'd feel much better if it had a bit more redundancy on the sides and back. Well, worry is too big a word as I don't really "worry" about it (I bought the AP2 eyes wide open), but it is a negative for AP2 in my mind. As said, my current thinking is that Tesla will likely add that redundancy in the next generation of AP hardware.
     
    • Like x 1
  14. shonline

    shonline Supporting Member

    Joined:
    Feb 25, 2016
    Messages:
    526
    Location:
    Mount Pleasant, SC
    Given their past responses to situations where customers were adversely impacted, I trust they will do the right thing if this occurs. If they do not, then it will mark the change from "On a Mission Tesla" to "just another greedy company driven by stock price Tesla".

    Ignoring the inflammatory nature of the use of "fanboy", it is interesting to see how the demographic of Tesla owners is changing. Many bought the car to support the cause, and now many are simply buying a car, with little to no investment in the cause. As such, expectations change and reactions to perceptions of being "lied to" and "ripped off" are becoming more common.

    This thread is a perfect example of this shift now occurring with the Tesla buyer - that will only increase when 3 is out.

    Inevitable, but a little sad nonetheless.
     
  15. calisnow

    calisnow Active Member

    Joined:
    Oct 11, 2014
    Messages:
    2,915
    Location:
    Los Angeles
    There's nothing massively expensive about a board swap.
     
  16. 182RG

    182RG Free The Service Manuals From Tyranny

    Joined:
    Apr 26, 2016
    Messages:
    723
    Location:
    Virginia
    Pedantic nit picking of words aside, there is no practical difference between the 2. There just isn't.
     
  17. 182RG

    182RG Free The Service Manuals From Tyranny

    Joined:
    Apr 26, 2016
    Messages:
    723
    Location:
    Virginia
    Hubris takes over.
     
    • Like x 1
  18. Reciprocity

    Reciprocity Active Member

    Joined:
    Feb 27, 2017
    Messages:
    4,164
    Location:
    Chicagoland
    I was thinking about this as well. 100,000 AI super computers re-acquired. What use could anyone possibly have for those? I know the boards are built to go into a car, but in theory, you could build a custom rack and drop these in to build a system capable of helping validate raw data coming back from the cars or for machine learning purposed to help the overall systems work. I know nVidia also supplies a server solution for the DrivePX total system. I dont know how much if it is used by Tesla or how much different those processors are. I am also guessing that the deal made with nVidia might have some outs based on promised performance.

    My guess is that way before they have produced the 500,000th Model 3, they will have their own chip sourced from Samsung or someone else. Similar to what Apple did when it first started building its own processors for the Iphone/Ipad.
     
  19. Reciprocity

    Reciprocity Active Member

    Joined:
    Feb 27, 2017
    Messages:
    4,164
    Location:
    Chicagoland
    1. The plan has always been fairly simple. Once the software is to a certain point, they can use the billions of miles of data they have to show regulators that the software is safer then a human. Also, regulators actually seem to be chomping at the bit to allow autonomous cars first. Many states seem to be stumbling over each other to allow autonomous driving so when there is a solution, they will expedite the acceptance vs sand bagging the process to slow it down.

    2. Lidar is not useful. The problem with Lidar + radar + vision is simple really. To much information in certain situations can actually be bad. What happens when your triple redundant system keeps coming up with different answers to whats in front of you? Which one do you trust? Elons made it very clear at the TED talk that the solution is a vision only solution. Radar would only be used to verify distance and to bounce under the front of cars in front of you to see what you cannot see. Lidar does not have a use in this solution becuase it cant bounce under anything and it cant see through rain or snow. Lidar will be used to create hi-def maps and to validate certain things on test cars, but not in production.

    You can either think Elon is nutz and everything he says is crazy and not to be trusted, or you can have a little faith and give the guy the benefit of the doubt as it relates to ultimately delivering things others said are impossible. Not just hard, but impossible. I choose to believe his is both nutz and going to deliver and I am just hanging as tight as I can for the ride.
     
    • Like x 1
  20. Reciprocity

    Reciprocity Active Member

    Joined:
    Feb 27, 2017
    Messages:
    4,164
    Location:
    Chicagoland
    I wish I knew what those words meant but I am getting the impression that you just insulted me.

    Here is the difference. The interviewer lacked the skills of a lawyer and asked the wrong question. And Elon still answered the right question of you listen/read the entire answer. Here are the questions and answers in context:

    It is clear that Elon is answering the question as it relates to sleeping in the car and feeling OK with not be conscious for the entirety of the trip and he clearly says that you wouldn't be fine if the system is 99.9% safe. He also says the system will mitigate the negative impact related to an accident. When its 99.9% safe and Elon can prove, using billions of miles of data to show how the car would have avoided things that humans did not, proving superhuman or 10x as safe as a human driver. 10x as safe does not mean you can wrap your head around sleeping in the car for the entire trip. But 10x as safe as a human is good enough to save hundreds of thousands of lives and that will get the attention of politicians who like votes from people who are alive, when they can get them. Dead people do in a pinch and even dogs/cats or whatever they can figure out.

    He clearly answers that it will take 2 years to get to the point where the average autonomous car crash that lead to some kind of injury to anyone would be 1 in a 100 or 1000 lifetimes. This is not required to save hundreds of thousands of lives, millions of lives over the next decade.

    Now if I were the interviewer, I would have asked when do you expect you will be able to start to seek approval to enable these features in HW2 cars that are on the road today? How long to you expect it to take?

    Elon being Elon would have said 6-12 months at most and that they can currently show something like 2x safer then a human and they just need more data to get to 10x.

    Now I need to figure out what "Pedantic nit picking" is as that sounds like something my wife does to me a lot.
     
    • Disagree x 1
    • Funny x 1

Share This Page

  • About Us

    Formed in 2006, Tesla Motors Club (TMC) was the first independent online Tesla community. Today it remains the largest and most dynamic community of Tesla enthusiasts. Learn more.
  • Do you value your experience at TMC? Consider becoming a Supporting Member of Tesla Motors Club. As a thank you for your contribution, you'll get nearly no ads in the Community and Groups sections. Additional perks are available depending on the level of contribution. Please visit the Account Upgrades page for more details.


    SUPPORT TMC