Comments by "Psiberzerker" (@Psiberzerker) on "The Real Moral Dilemma of Self-Driving Cars" video.

"How much should you trust an autonomous car?" Honestly, that's not the moral dilemma here. How much do you trust YOURSELF to drive the car? You don't trust the car, which was designed in this case by BMW, but you don't even ask yourself how much you can be trusted behind the wheel. Okay, you can handle it. (Every drunk driver ever) but what about that guy? What about that guy "Fixing" his autonomous car, so it takes the route He wants to take, because he knows the best way to work? Not trying to mess up the navigation system that keeps his car from navigating into your car, but how much do you trust Bob to mess with his autonomous car? The car isn't the problem.
3
Cars don't kill people, drivers kill people. Incidentally, 1:47 How many people are killed by guns, every single year? Pick a year, any year, look up the fatalities for Guns. Pistols, Rifles, Shotguns. Now, look at fatalities from Automobiles. Cars, Trucks, and Motorcycles. Accidents, Suicides, and Murders. Taken together, cherry pick the best year you can find where guns are significantly deadlier than cars, or vice-versa. It's kinda funny how we're even having this Gun Debate, but people don't "Trust" a car they can't drive. "If it kills someone, I'm legally responsible." Yeah, but if you kill someone, you're legally, and Morally responsible.
1
@sage1476 Okay, how much do you trust Windows after an update? It's not just the software, how much brake fluid did Bob drain yanking stuff out of the car before he got down to the CPU? How many sensors did he disable before he tried to put it all back? We humans are not capable of writing the system that can compensate for other Humans.
1
@sage1476 Apollo 1, Apollo 13, Challenger, and Colombia were pretty much capable of running their missions with minimal pilot input, until landing. (Apollo 1 wasn't even supposed to leave the ground.) They all failed because of bad components. That's basically the level of human-computer-machine interface. Pretty much the most reliable component in that loop is the auto-pilot. Air disasters are almost invariably either operator Eror, or bad maintenance. They're almost never the auto-pilot.
1
@sage1476 Yes, but it's still, literally Rocket Science, and the Autopilot/telemetry never failed. "Oh, it's only 35 year old" rocket science. So, it's still not impressive, or 7 of the most highly trained people in America trusting a self piloted space-plane. So, we had similar technology that was virtually 100% reliable, decades ago. But people don't trust a car they can't control. Or, people just like being in control.
1
An autonomous car won't work. On our roads, that's great, you can create a car that won't hit other cars on the street. You can't make a car that won't hit a child in the street. That's what we're talking about here, all the times where a computer has to make a choice, which human being to kill (In milliseconds. I'd trust an onboard computer to make that choice in milliseconds, because it can think in milliseconds) but our roads weren't designed for them. They're designed for people, that have to get from point A, to point B, even when point B is right across the street from Point A. Or point B is bouncing across the street, because Suzy didn't catch it.
1