4 ms·

I think most of the criticism with self driving is around the over-hyped under-delivered "full self driving" from Musk -- it actually seems to be getting worse

I think most of the criticism with self driving is around the over-hyped under-delivered "full self driving" from Musk -- it actually seems to be getting worse compared to 5 years ago. Apparently Cruise and Waymo are doing much more impressive things, but they only get a fraction of the press.

benlivengood 4y ago
I'm curious how many people who criticize full self drive have actually ridden in the car while it navigates city streets. It makes about as many mistakes as a 1-3 years of practice new driver. I don't own a Tesla so I've ridden with friends. It's a barely noticeable difference between human and FSD operation, with very few overrides necessary.
chrisco255 4y ago
If they're doing more impressive things why is it not in production at scale, right now? Just earlier today there was a story of a Waymo car getting stuck in an intersection in SF. Those cars still need constant intervention and human assistance that's why.
- daveguy 4y ago
  The miles between intervention is known for Waymo and Cruise. Tesla refuses to publish this data.
  - 93po 4y ago
    Waymo and Cruise have published their known operating areas. It's tiny and only during certain times of day (middle of the night). Tesla operates nearly everywhere. The scale of what's happening isn't even comparable.
    - daveguy 4y ago
      Estimates from users of FSD are ~5 miles between disengagements. https://sites.google.com/view/fsdbetacommunitytracker/home?pli=1 https://sites.google.com/view/fsdbetacommunitytracker/home?p... Even the best self reported areas are less than 100 miles between disengagement. Waymo and Cruise are at around 30,000 miles between disengagement. Tesla has released nothing.
      - 93po 4y ago
        Tesla has FSD released on literally every car in the US that has purchased it. Hundreds of thousands of cars. No one serious in the self driving space is using disengagements to pick winners or leaders on the path to commercialization. There is no standard about what qualifies as a disengagement and companies will interpret and report their disengagements differently. Those interpretations also change over time, making it even more difficult to actually understand how a company has progressed in its technology. And again you're missing the point. Waymo/Cruise are operating in an extremely different and much more narrow context than Tesla's FSD. Let's go put Waymo on a random road in a random city and see how it fares compared to Tesla FSD. I guarantee it will be tremendously worse and no one who knows anything about the space would argue otherwise.
        daveguy 4y ago
        Hundreds of thousands of 5 mile jaunts. Also, why is Tesla so adamant about how many accidents occured simply with FSD disabled? Shouldn't we know how long prior to the accident FSD was engaged? I mean, hitting the brakes in panic only technically makes it "not FSD". I'd be happy if you showed me any reported data, self or Tesla, of 30k miles before disengagement with FSD. Even within one order of magnitude instead of two.
        93po 4y ago
        Show me data of Waymo or Cruise doing 30k miles before disengagement in the 99% of the country where they don't operate. I guarantee their stats are worse anywhere outside their carefully chosen, extremely narrow operating areas.
93po 4y ago
> it actually seems to be getting worse compared to 5 years ago. You pretty clearly don't know what you're talking about. > Apparently Cruise and Waymo are doing much more impressive things, but they only get a fraction of the press. They are doing objectively much less impressive things by relying on sensors that require much less intelligent analysis and will lead them to dead ends at the higher end of abilities that Tesla's approach won't.
- daveguy 4y ago
  They can't even keep up with self-parking technology. All the other electric car manufacturers don't seem to have any problem with it. But Tesla's implementation is laughable. https://m.youtube.com/watch?v=nsb2XBAIWyA https://m.youtube.com/watch?v=nsb2XBAIWyA

benlivengood 4y ago

I'm curious how many people who criticize full self drive have actually ridden in the car while it navigates city streets. It makes about as many mistakes as a 1-3 years of practice new driver. I don't own a Tesla so I've ridden with friends. It's a barely noticeable difference between human and FSD operation, with very few overrides necessary.

chrisco255 4y ago

If they're doing more impressive things why is it not in production at scale, right now? Just earlier today there was a story of a Waymo car getting stuck in an intersection in SF. Those cars still need constant intervention and human assistance that's why.

daveguy 4y ago
The miles between intervention is known for Waymo and Cruise. Tesla refuses to publish this data.
- 93po 4y ago
  Waymo and Cruise have published their known operating areas. It's tiny and only during certain times of day (middle of the night). Tesla operates nearly everywhere. The scale of what's happening isn't even comparable.
  - daveguy 4y ago
    Estimates from users of FSD are ~5 miles between disengagements. https://sites.google.com/view/fsdbetacommunitytracker/home?pli=1 https://sites.google.com/view/fsdbetacommunitytracker/home?p... Even the best self reported areas are less than 100 miles between disengagement. Waymo and Cruise are at around 30,000 miles between disengagement. Tesla has released nothing.
    - 93po 4y ago
      Tesla has FSD released on literally every car in the US that has purchased it. Hundreds of thousands of cars. No one serious in the self driving space is using disengagements to pick winners or leaders on the path to commercialization. There is no standard about what qualifies as a disengagement and companies will interpret and report their disengagements differently. Those interpretations also change over time, making it even more difficult to actually understand how a company has progressed in its technology. And again you're missing the point. Waymo/Cruise are operating in an extremely different and much more narrow context than Tesla's FSD. Let's go put Waymo on a random road in a random city and see how it fares compared to Tesla FSD. I guarantee it will be tremendously worse and no one who knows anything about the space would argue otherwise.
      - daveguy 4y ago
        Hundreds of thousands of 5 mile jaunts. Also, why is Tesla so adamant about how many accidents occured simply with FSD disabled? Shouldn't we know how long prior to the accident FSD was engaged? I mean, hitting the brakes in panic only technically makes it "not FSD". I'd be happy if you showed me any reported data, self or Tesla, of 30k miles before disengagement with FSD. Even within one order of magnitude instead of two.
        93po 4y ago
        Show me data of Waymo or Cruise doing 30k miles before disengagement in the 99% of the country where they don't operate. I guarantee their stats are worse anywhere outside their carefully chosen, extremely narrow operating areas.

93po 4y ago

> it actually seems to be getting worse compared to 5 years ago. You pretty clearly don't know what you're talking about. > Apparently Cruise and Waymo are doing much more impressive things, but they only get a fraction of the press. They are doing objectively much less impressive things by relying on sensors that require much less intelligent analysis and will lead them to dead ends at the higher end of abilities that Tesla's approach won't.

daveguy 4y ago
They can't even keep up with self-parking technology. All the other electric car manufacturers don't seem to have any problem with it. But Tesla's implementation is laughable. https://m.youtube.com/watch?v=nsb2XBAIWyA https://m.youtube.com/watch?v=nsb2XBAIWyA