3 ms·

I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. It is why humans are

I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. It is why humans are simultaneously great and terrible at driving. One slightly bad driver can be accounted for by several slightly better drivers navigating around the bad one's behavior. This smooth "self-correction" on the roadway leads to lots of rule breaking but also probably saves a lot of time/lives by allowing individual humans to navigate a situation (especially since lots of people forget the exact rules of the road...ex. do you yield in a turning situation or not, what exceptions would you not yield?) However, I really do like your last point, the easiest way to solve is just remove the cars from as much of the equation and introduce transit that follows rules (trains, busses -higher compliance with road rules perhaps) and let humans navigate the last meter themselves

dionidium 5y ago
> I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. This makes more sense as an objection in a vacuum than it does in the reality where the alternative is that speeding is a big problem within cities. We don't have to solve every edge case. We just need the edge cases to be less of an issue than speeding already is. And that's a much easier bar. Asking, "what if we do this?" is fine, but one should also always consider, "what if we don't?"
- darkwizard42 5y ago
  But that is the point... if we don't do something here, humans can continue to take their own actions to resolve the situation (speed up with traffic or move to the right or choose an alternate route). That level of flexibility allows for each person to make the choice they want and feel comfortable with (as opposed to the car making a decision for you)
  - dionidium 5y ago
    Let's be clear here that "letting the car make the decision for you" in context means "you can't go as fast as you want in an urban area because the car will prevent it." As in, you can't drive many multiples of the speed limit in urbanized areas packed with buildings and pedestrians and other cars. The car should prevent that in practically all cases. The car does know that that's a good idea in practically every conceivable case. A human driver who thinks they should drive many multiples of the speed limit in an urban area is wrong. It's so manifestly wrong that it shouldn't even be technically possible. There's no good reason to manufacture a car that responds so stupidly to such a dangerous and irresponsible input. And when the day finally comes that the conditions are just right so that a driver dies attempting to flee an avalanche because this tech prevented them from escaping at speed, we will put a lone checkmark in the "lives lost to this technology" column that's adjacent to a "lives saved" file 10 miles long. If I could be permitted just a little bit of levity: https://twitter.com/dril/status/464802196060917762?lang=en https://twitter.com/dril/status/464802196060917762?lang=en
shagie 5y ago
> I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. There's a story out there by Cory Doctorow about self driving cars. Car Wars - https://craphound.com/news/2016/11/23/car-wars-a-dystopian-science-fiction-story-about-the-nightmare-of-self-driving-cars/ https://craphound.com/news/2016/11/23/car-wars-a-dystopian-s... Its no longer accessible on the site... but you can find it on wayback. https://web.archive.org/web/20190225032830/https://this.deakin.edu.au/self-improvement/car-wars https://web.archive.org/web/20190225032830/https://this.deak... (though the formatting in the twitter feed like section is a bit weird) The part that is most applicable to this begins at "But it was more complicated than just biased data".

dionidium 5y ago

> I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. This makes more sense as an objection in a vacuum than it does in the reality where the alternative is that speeding is a big problem within cities. We don't have to solve every edge case. We just need the edge cases to be less of an issue than speeding already is. And that's a much easier bar. Asking, "what if we do this?" is fine, but one should also always consider, "what if we don't?"

darkwizard42 5y ago
But that is the point... if we don't do something here, humans can continue to take their own actions to resolve the situation (speed up with traffic or move to the right or choose an alternate route). That level of flexibility allows for each person to make the choice they want and feel comfortable with (as opposed to the car making a decision for you)
- dionidium 5y ago
  Let's be clear here that "letting the car make the decision for you" in context means "you can't go as fast as you want in an urban area because the car will prevent it." As in, you can't drive many multiples of the speed limit in urbanized areas packed with buildings and pedestrians and other cars. The car should prevent that in practically all cases. The car does know that that's a good idea in practically every conceivable case. A human driver who thinks they should drive many multiples of the speed limit in an urban area is wrong. It's so manifestly wrong that it shouldn't even be technically possible. There's no good reason to manufacture a car that responds so stupidly to such a dangerous and irresponsible input. And when the day finally comes that the conditions are just right so that a driver dies attempting to flee an avalanche because this tech prevented them from escaping at speed, we will put a lone checkmark in the "lives lost to this technology" column that's adjacent to a "lives saved" file 10 miles long. If I could be permitted just a little bit of levity: https://twitter.com/dril/status/464802196060917762?lang=en https://twitter.com/dril/status/464802196060917762?lang=en

shagie 5y ago

> I think the issue of "always bringing up edge cases" is meant to illustrate that a LOT can go wrong with some of these hard and fast rules. There's a story out there by Cory Doctorow about self driving cars. Car Wars - https://craphound.com/news/2016/11/23/car-wars-a-dystopian-science-fiction-story-about-the-nightmare-of-self-driving-cars/ https://craphound.com/news/2016/11/23/car-wars-a-dystopian-s... Its no longer accessible on the site... but you can find it on wayback. https://web.archive.org/web/20190225032830/https://this.deakin.edu.au/self-improvement/car-wars https://web.archive.org/web/20190225032830/https://this.deak... (though the formatting in the twitter feed like section is a bit weird) The part that is most applicable to this begins at "But it was more complicated than just biased data".