5 ms·

I guess until all autonomous vehicles are fully equipped with real sensors (likely both lidar and radar) so that cars can almost-mechanically stop before forese

by esjeon 3y ago

I guess until all autonomous vehicles are fully equipped with real sensors (likely both lidar and radar) so that cars can almost-mechanically stop before foreseeable collisions without worrying about AI hiccups. This is technically possible even now, but it's way too expensive for now.

jvanderbot 3y ago
Unless a lot of money is paid out in campaign contributions, I suspect you're right. I'm worried about a time when the _statistical_ safety of AI driving is better than humans across the board, but there are horrifying correlations between AI induced accidents, like "Children on tricycles are not well recognized" or "slight fog in Seattle induces 150 car pileup". Or, AI causes worse accidents less frequently than humans. Whereas humans have lots more fender benders and minor injuries to older adults. Not saying this is true, but it's a scenario I think about a lot.
- brandonagr2 3y ago
  A lot more minor injuries? You are aware that over 100 people are killed in cars every single day in the US alone right?
  - colonelpopcorn 3y ago
    TBH, that stat seems low.
  - jvanderbot 3y ago
    I think you missed the part where I'm not saying this is current reality, just saying it's something I worry about. Continuing your example just for exposition / hypothetical, if those are 100 individual deaths with varying causes, vs 99 children run over by AI b/c they were walking in a line holding hands and were interpreted as road markers, well that's not _better_ is it? This is purely hypothetical, but statistics gloss over those human details.
- richiebful1 3y ago
  Given how autonomous vehicles drive today (cautiously, within the speed limit), I feel pretty good about autonomous vehicles easily beating out humans on highway driving. I'd love to see a highway-oriented true full self driving mode developed, and I'd probably pay an additional ~10-20k to get it
  - toyg 3y ago
    Highways/motorways should be the perfect ground for "self-driving trains" of cars: self-regulating convoys of vehicles going at similar speeds between exits, "attached" at safe distances. It would dramatically reduce stress and fatigue (and hence accidents) among professional drivers and other heavy users. Particularly in the US, it's an absolute no-brainer. The problem is finding who will pay for the necessary infrastructure.
    - jvanderbot 3y ago
      Highway self driving is already here. Any new car has lane-keeping and adaptive cruise control. That's all you need. For fair weather anyway.
brandonagr2 3y ago
Why should lidar be a requirement when that's not a requirement for humans to be able to drive? The goal is to reduce traffic fatalities, humans are not reliable, if you say we can't have automation until its 100% perfect, then you are actively choosing to have more traffic fatalities.
- unshavedyak 3y ago
  Depends, is Tesla currently as good as Waymo? I don't think the OP was arguing for perfection, i think they were recognizing the current state of everything. Our ability to make AI as good as human with as limited sensors as humans is not great, currently. Throw in more sensors though, and you can offset deficiencies in human-visible-only AI detection. To me, more sensors means faster implementation. Quicker to get on roads. Less total deaths. At least in theory, and in what successes i'm seeing.
- edgyquant 3y ago
  What else are you planning to use to scan the environment? Why should we not require multiple instruments?
- lawlessone 3y ago
  >when that's not a requirement for humans to be able to drive? If it get's cheap enough it will be. But the AI isn't human and things a human would recognize, like a a large grey truck up ahead doing a turn on an overcast day an AI can just miss. > if you say we can't have automation until its 100% perfect, That's why we need the lidar.
- timthelion 3y ago
  This is a bad argument because humans use two eyes that move constantly in unison, have massively higher dynamic range than any digital camera, and can accurately measure distance in a fraction of a second by changing their focus distance. A web cam really can't replace human sight, if it was two very high speed camersa that moved like human eyes than your argument would stand, but it is not!
- michaelt 3y ago

by esjeon 3y ago

jvanderbot 3y ago

Unless a lot of money is paid out in campaign contributions, I suspect you're right. I'm worried about a time when the _statistical_ safety of AI driving is better than humans across the board, but there are horrifying correlations between AI induced accidents, like "Children on tricycles are not well recognized" or "slight fog in Seattle induces 150 car pileup". Or, AI causes worse accidents less frequently than humans. Whereas humans have lots more fender benders and minor injuries to older adults. Not saying this is true, but it's a scenario I think about a lot.

brandonagr2 3y ago
A lot more minor injuries? You are aware that over 100 people are killed in cars every single day in the US alone right?
- colonelpopcorn 3y ago
  TBH, that stat seems low.
- jvanderbot 3y ago
  I think you missed the part where I'm not saying this is current reality, just saying it's something I worry about. Continuing your example just for exposition / hypothetical, if those are 100 individual deaths with varying causes, vs 99 children run over by AI b/c they were walking in a line holding hands and were interpreted as road markers, well that's not _better_ is it? This is purely hypothetical, but statistics gloss over those human details.
richiebful1 3y ago
Given how autonomous vehicles drive today (cautiously, within the speed limit), I feel pretty good about autonomous vehicles easily beating out humans on highway driving. I'd love to see a highway-oriented true full self driving mode developed, and I'd probably pay an additional ~10-20k to get it
- toyg 3y ago
  Highways/motorways should be the perfect ground for "self-driving trains" of cars: self-regulating convoys of vehicles going at similar speeds between exits, "attached" at safe distances. It would dramatically reduce stress and fatigue (and hence accidents) among professional drivers and other heavy users. Particularly in the US, it's an absolute no-brainer. The problem is finding who will pay for the necessary infrastructure.
  - jvanderbot 3y ago
    Highway self driving is already here. Any new car has lane-keeping and adaptive cruise control. That's all you need. For fair weather anyway.

brandonagr2 3y ago

Why should lidar be a requirement when that's not a requirement for humans to be able to drive? The goal is to reduce traffic fatalities, humans are not reliable, if you say we can't have automation until its 100% perfect, then you are actively choosing to have more traffic fatalities.

unshavedyak 3y ago
Depends, is Tesla currently as good as Waymo? I don't think the OP was arguing for perfection, i think they were recognizing the current state of everything. Our ability to make AI as good as human with as limited sensors as humans is not great, currently. Throw in more sensors though, and you can offset deficiencies in human-visible-only AI detection. To me, more sensors means faster implementation. Quicker to get on roads. Less total deaths. At least in theory, and in what successes i'm seeing.
edgyquant 3y ago
What else are you planning to use to scan the environment? Why should we not require multiple instruments?
lawlessone 3y ago
>when that's not a requirement for humans to be able to drive? If it get's cheap enough it will be. But the AI isn't human and things a human would recognize, like a a large grey truck up ahead doing a turn on an overcast day an AI can just miss. > if you say we can't have automation until its 100% perfect, That's why we need the lidar.
timthelion 3y ago
This is a bad argument because humans use two eyes that move constantly in unison, have massively higher dynamic range than any digital camera, and can accurately measure distance in a fraction of a second by changing their focus distance. A web cam really can't replace human sight, if it was two very high speed camersa that moved like human eyes than your argument would stand, but it is not!
michaelt 3y ago