Originally Posted by
genec
OK. Let's make some assumptions to test my THEORY.
Let's say we finally get the AI nearly right and it can do Level 5 Self Driving... at what point is it acceptable as a replacement for human drivers?
- Would you accept that AVs (with AI) are superior to human drivers if the number of traffic deaths was cut in half... to say merely 20,000 a year?
- How about if AVs were 90% perfect, and the number of traffic deaths was reduced to a mere 4000 a year?
- Or do you demand 99% perfection of AI, and only 400 deaths a year.
- Or will you only accept 100% perfection of AI... and no traffic deaths?
So what is "good enough?" 1, 2, 3 or only 4... and why? Considering that right now we "allow" for 40,000 deaths by drivers on our highways each year?
If it actually made people safer (not just in terms of numbers of deaths, btw), then of course it should be adopted. You made up all that crap about demanding 99% perfection (whatever the hell that means) out of your own head--that's your strawman. No idea why a 99% reduction in deaths translates to 99% perfection, but that's your inane terminology, not mine.
All of this is irrelevant, because there's absolutely no reason to assume your numbers have anything to do with reality. And of course the 50% reduction in deaths would be a good thing, who besides you said it wouldn't be?
So tell me, what is your evidence for the proposition that people are demanding "100% perfection"? That's the kind of strawman people who can't defend the current state of the technology as currently being safe for road use would throw out--it's a "look at the birdie" distraction from you having no real argument.. You didn't really think I was going to say that the number of deaths shouldn't be reduced, did you?