That's the thing... We don't know because there is no regulatory body in place overseeing this testing. The proper way to do this is to set up a system where by these events are well understood and they aren't. So until that is in place, self driving cars shouldn't be in the public at all.
This is going to change a lot of self driving conversations from here on out. No longer can self driving fans claim that the cars are faster than a human, or can see better and make out shapes better than a human. This still may be true if a theoretical state of perfection is reached between correct (expensive) sensors and programming but there is such a long way to go. Show me an AI that can see a dark mass of shapes and pick out the human or moose standing still. Show me the AI that recognizes every type of throw toy so that it may anticipate someone running out after it into the road. This case only emphasizes the fact that it will be requires, and that the people who were dubious of current technology were right.
The problem is:
1) It is damn expensive to do it right, and
2) There is no standard method to prove if someone is doing it right before allowing them on public roads
In the grand scheme of things, it doesn't matter. It's this "she was doing something stupid, a car shouldn't have to deal with that" attitude why there are so many accidents in the US in the first place. You don't drive faster than you can see. It's that simple, no matter who or what is driving.
It doesn't change the fact that all drivers should drive defensively, even automated ones. A lot of defensive driving is about expecting people to do stupid things... because a lot of them are stupid. And anyway, it could have been a fairly large animal on the road; large enough to injure the driver. You don't drive with a longer stopping distance than your headlights. You keep your eyes open. Both car and human failed spectacularly at this. Automated cars aren't even supposed to *need* headlights.
She was standing on the side of the road for the full approach. Headlights work for over 100 feet, the car should definitely not been speeding past the capability of the headlights.. You can't trust a video to show human eye accurate details in darkness. Automated cars are supposed to have sensors that are not defeated by darkness.
But... but.... They can see in the dark!... ok never mind that one. But they have amazing reaction time! Ok never mind that one too. Ok I guess you're right.
Uber still has to be able to identify from the images someone standing still by the side of the road that is preparing to walk into it. I'm not convinced they can do that with any sensor.
Why the fuck does the pedestrian have to be moving? A pedestrian can walk out into the road after not moving just as easily; the car needs to sense the danger and slow down regardless.
No one is saying that there aren't situations that sensors are much better in. What remains to be seen is if that will ever outweigh the situations that they suck at.
Slashdot Mirror
That's the thing... We don't know because there is no regulatory body in place overseeing this testing. The proper way to do this is to set up a system where by these events are well understood and they aren't. So until that is in place, self driving cars shouldn't be in the public at all.
This is going to change a lot of self driving conversations from here on out. No longer can self driving fans claim that the cars are faster than a human, or can see better and make out shapes better than a human. This still may be true if a theoretical state of perfection is reached between correct (expensive) sensors and programming but there is such a long way to go. Show me an AI that can see a dark mass of shapes and pick out the human or moose standing still. Show me the AI that recognizes every type of throw toy so that it may anticipate someone running out after it into the road. This case only emphasizes the fact that it will be requires, and that the people who were dubious of current technology were right.
The problem is:
1) It is damn expensive to do it right, and
2) There is no standard method to prove if someone is doing it right before allowing them on public roads
It's not up to Uber or any self driving company to make stupidity a capital offense.
In the grand scheme of things, it doesn't matter. It's this "she was doing something stupid, a car shouldn't have to deal with that" attitude why there are so many accidents in the US in the first place. You don't drive faster than you can see. It's that simple, no matter who or what is driving.
It doesn't change the fact that all drivers should drive defensively, even automated ones. A lot of defensive driving is about expecting people to do stupid things... because a lot of them are stupid. And anyway, it could have been a fairly large animal on the road; large enough to injure the driver. You don't drive with a longer stopping distance than your headlights. You keep your eyes open. Both car and human failed spectacularly at this. Automated cars aren't even supposed to *need* headlights.
Sorry I should have used a
She was standing on the side of the road for the full approach. Headlights work for over 100 feet, the car should definitely not been speeding past the capability of the headlights.. You can't trust a video to show human eye accurate details in darkness. Automated cars are supposed to have sensors that are not defeated by darkness.
It's kind of funny that you need to explain to people that cars should avoid hitting pedestrians.
But... but.... They can see in the dark!... ok never mind that one. But they have amazing reaction time! Ok never mind that one too. Ok I guess you're right.
It is a widely held belief that self driving was Uber's ulterior motive all along. They lose shitloads of money with people driving.
Uber still has to be able to identify from the images someone standing still by the side of the road that is preparing to walk into it. I'm not convinced they can do that with any sensor.
Why the fuck does the pedestrian have to be moving? A pedestrian can walk out into the road after not moving just as easily; the car needs to sense the danger and slow down regardless.
Why does she have to move for the car to see her? The car should sense a dangerous situation whether there is movement or not.
There are actually a couple large testing facilities for this.
Well there you go. Clearly these cars should be kept on the road so Uber has every opportunity to make their technology better.
Bingo, you also found the cutoff point at which they should be on public roads.
It needs to be better than a 95 year old with cataracts!
So a car that hits ladies with bikes because it can't see them is on track to be an 'improvement'?
So if it's literally not right in front of your eyes, your brain won't really process it very well if at all.
Uh, that's why we have 'necks' ?
No one is saying that there aren't situations that sensors are much better in. What remains to be seen is if that will ever outweigh the situations that they suck at.
Really? You think hitting ladies with bikes in the middle of an open road is the path to success here?
Future victims of autonomous car accidents thank you for the abysmally low bar you put on safety levels for autonomous cars. Uber investors thank you.
I was actually thinking they should have a person ask them questions about the road ahead every ten minutes or so. That may keep them more involved.
To be successful, automated cars will need to make out things much smaller than adult women pushing bikes.