Advertisement
If you have a new account but are having problems posting or verifying your account, please email Niamh on [email protected] for help. Thanks :)
New AMA with a US police officer (he's back!). You can ask your questions here

Hyperchange buys Hypership, is the cat out of the bag with Autopilot?

  • 02-09-2020 2:42pm
    #1
    Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    Fast forward to 17 minutes in, do you agree that autopilot will make current teslas look silly cheap when autopilot goes autonomous..



Comments

  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    "when"? If.


  • Moderators, Motoring & Transport Moderators, Regional East Moderators Posts: 6,084 Mod ✭✭✭✭ liamog


    If autopilot is as successful as Tesla likes to speculate, they won't sell any vehicles. Why would you a sell a device to a consumer that cost's €40,000 and is capable of making €200,000 as a robotaxi.


  • Registered Users Posts: 17,907 ✭✭✭✭ ELM327


    The "robotaxis" will not be sold with that earning potential at that price, that's for sure.


  • Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    exactly .. when the cat IS out of the bag a new tesla with autonomy will cost much much more .. making the current ones (that have autonomy) look silly cheap.

    anyone getting FOMO ( fear of missing out ) ?


  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    Well contrary to what Gali "what up" Russell asserts in his almost unwatchable stock pumping drivel, buying FSD is not taking an option on future fully autonomous functions, because the option expires with the car, and cars have an extremely short life relative to the glacial progress of AI.

    The battery will be toast long before the autonomy works.

    So no FOMO.


  • Advertisement
  • Moderators, Category Moderators, Motoring & Transport Moderators, Regional East Moderators Posts: 58,865 CMod ✭✭✭✭ unkel


    There's nothing new here, is there?

    Musk has been saying for a long time that FSD will get more expensive as it improves. I didn't believe him, when the company wasn't doing too well, I reckoned he might have to sell it at a discount at some point in the future, as it is like an easy way to boost revenue, with zero costs. I might turn out to be right yet, but it doesn't look like it

    Musk also said that for someone buying a new car, it was "insane not to buy a Tesla". He might yet prove himself right although he obviously has always been extremely optimistic. Did make him one of the richest men on the planet though. Wouldn't bet against him

    If I'd had a bit more faith in him, I would have added a very significant chunk to my pension fund. Alas.

    "Wind is Ireland's oil" - An Taoiseach, 25/05/2022



  • Registered Users Posts: 6,396 ✭✭✭ jhegarty


    There is a very big difference between FSD that will work 99.9% of the time (required for normal people to be happy) and the FSD that will work 99.999% of the time (required for robotaxis).

    As an example , no one would say FSD needs to drive the wrong way up the hard-shoulder of the motorway when directed by a workman with a flag and yellow vest. But a robotaxi needs to handle that.


  • Moderators, Motoring & Transport Moderators, Regional East Moderators Posts: 6,084 Mod ✭✭✭✭ liamog


    I now have the image of an FSD Tesla as a petulant toddler. It's fine most of the time, but when it gets to something it doesn't like it just throws a strop and stops at the side of the road.
    I guess that could be solved by remote operation for a robotaxi fleet


  • Registered Users Posts: 8,830 ✭✭✭ markpb


    liamog wrote: »
    I now have the image of an FSD Tesla as a petulant toddler. It's fine most of the time, but when it gets to something it doesn't like it just throws a strop and stops at the side of the road.

    Well they already do that from time to time, just not on the side of the road :)


  • Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    jhegarty wrote: »
    There is a very big difference between FSD that will work 99.9% of the time (required for normal people to be happy) and the FSD that will work 99.999% of the time (required for robotaxis).

    As an example , no one would say FSD needs to drive the wrong way up the hard-shoulder of the motorway when directed by a workman with a flag and yellow vest. But a robotaxi needs to handle that.

    I heard a solution proposed (from waymo I think) for these outliers would be for a small team of humans back at base. If the car was confused it would simply stop and call home .. the team would then remotely drive the car slowly around the issue, these would then feedback to the net and reduce the outliers over time.


  • Advertisement
  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    celtic_oz wrote: »
    I heard a solution proposed (from waymo I think) for these outliers would be for a small team of humans back at base. If the car was confused it would simply stop and call home .. the team would then remotely drive the car slowly around the issue, these would then feedback to the net and reduce the outliers over time.

    ...which is fine unless you happen to have roadworks in a mobile data dead spot like the N11 at Kilmacaogue, and then you end up with a swarm of stranded Teslas creating total chaos.

    I'm not sure whether the latency on Starlink is low enough to remotely pilot a car safetly.


  • Moderators, Category Moderators, Motoring & Transport Moderators, Regional East Moderators Posts: 58,865 CMod ✭✭✭✭ unkel


    jhegarty wrote: »
    There is a very big difference between FSD that will work 99.9% of the time (required for normal people to be happy) and the FSD that will work 99.999% of the time (required for robotaxis)

    That depends on what percentage a human driver is on average. FSD does not have to be perfect. Once FSD is significantly better than a human driver (this can be fairly easily proven based on say fatal accidents / 100m km), it is a done deal and authorities will approve it. Or in Musk's words: people will actually demand it.

    "Wind is Ireland's oil" - An Taoiseach, 25/05/2022



  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    unkel wrote: »
    That depends on what percentage a human driver is on average. FSD does not have to be perfect. Once FSD is significantly better than a human driver (this can be fairly easily proven based on say fatal accidents / 100m km), it is a done deal and authorities will approve it. Or in Musk's words: people will actually demand it.

    Zero deaths is trivially achieved by setting the maximum speed to below that which causes fatal accidents in head-on collisions.

    Safety is not the sole requirement of driving, and by extension autonomous driving.


  • Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    Lumen wrote: »
    ...which is fine unless you happen to have roadworks in a mobile data dead spot like the N11 at Kilmacaogue, and then you end up with a swarm of stranded Teslas creating total chaos.

    I'm not sure whether the latency on Starlink is low enough to remotely pilot a car safetly.

    starlink latency will be good enough link


  • Closed Accounts Posts: 979 ✭✭✭ Thierry12


    liamog wrote: »
    Why would you a sell a device to a consumer that cost's €40,000 and is capable of making €200,000 as a robotaxi.

    Genius Liamog :pac:

    I can buy a robo lawnmower for €1,000

    If I cut a few people's grass once a week for €5 a go, I should be able to make €5,000 after a few years

    That lawnmower shoud be costing way more than €1,000


  • Moderators, Motoring & Transport Moderators, Regional East Moderators Posts: 6,084 Mod ✭✭✭✭ liamog


    Thierry12 wrote: »
    Genius Liamog :pac:

    I can buy a robo lawnmower for €1,000

    If I cut a few people's grass once a week for €5 a go, I should be able to make €5,000 after a few years

    That lawnmower shoud be costing way more than €1,000

    You'd spend more money than you make driving the robo mower to the customers house and having to drop the guide wire around the lawn. Wouldn't be worth the effort.


  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    liamog wrote: »
    You'd spend more money than you make driving the robo mower to the customers house and having to drop the guide wire around the lawn. Wouldn't be worth the effort.

    Also, robomowers typically are specced for maybe 50% utilisation on a single lawn, so even if it could fly between houses autonomously it wouldn't be much more valuable.

    Regardless, arguing that a Tesla is worth 200k because robotaxis completely ignores commoditisation effects. Robotaxis won't be competing against humans, they'll be competing against other robotaxis.


  • Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    Lumen wrote: »
    Robotaxis won't be competing against humans, they'll be competing against other robotaxis.

    yes all those leafs, i3's, zoe's etc etc are going to magically turn into robo-taxis.. oh no wait


    Tesla could be largely competition free for 5 years if they get the jump on this.


  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    celtic_oz wrote: »
    yes all those leafs, i3's, zoe's etc etc are going to magically turn into robo-taxis.. oh no wait


    Tesla could be largely competition free for 5 years if they get the jump on this.

    Why do robitaxis need to be electric? Why does the car brand need to own or develop the tech?

    Tesla is not leading autonomous driving tech. It just has a massive fleet of cars with fancy cruise control.


  • Registered Users Posts: 1,085 ✭✭✭ celtic_oz


    Lumen wrote: »
    Why do robitaxis need to be electric? Why does the car brand need to own or develop the tech?

    Tesla is not leading autonomous driving tech. It just has a massive fleet of cars with fancy cruise control.

    They will make more profit if electric

    There is no other company that even CLAIMS it will have millions of cars on the road that can magically flip into robotaxis OVERNIGHT.

    Tesla COULD be largely competition free for 5 years IF they get the jump on this.


  • Advertisement
  • Registered Users Posts: 30,514 ✭✭✭✭ Lumen


    celtic_oz wrote: »
    They will make more profit if electric

    There is no other company that even CLAIMS it will have millions of cars on the road that can magically flip into robotaxis OVERNIGHT.

    Tesla COULD be largely competition free for 5 years IF they get the jump on this.

    OK, at least I understand your argument now. :D

    Practically speaking...I have no desire to operate a taxi service so if this became possible (and I had bought FSD, which I haven't, because it's rubbish) I would just sell my Tesla to a robotaxi company and buy a Taycan.

    However, anyone buying a Tesla with FSD on that basis is an optimist/idiot, because whilst the software engineers spend years rewriting and debugging, and the lawyers spend years convincing governments to make it legal (whilst their competitors spend vast money on stalling that process) the car is depreciating to nothing.

    In any case, as a likely enthusiastic consumer of robotaxi services I have no interest in squeezing into the front or back of a Model 3, I want something more like a Canoo. Based on Tesla's historical time to market, it would take them several years to build one and then several more years to make the suicide doors work.

    5d8b976c2e22af376850c3cf?width=1100&format=jpeg&auto=webp


Advertisement