I think I agree with most of what you are saying here. Definitely with the part where we intuitively solve complex nonlinear differential equations tailored to specific classes of problems that do not necessarily generalize well outside their domain, and do not rise to the level of conscious thought. The centipede dilemma seems topical here.
As for one of your examples, autonomous driving, my guess is that computers are better at it than most human drivers in 99% of all cases, but the complexity of the remaining 1% is what kills it.
I think I agree with most of what you are saying here. Definitely with the part where we intuitively solve complex nonlinear differential equations tailored to specific classes of problems that do not necessarily generalize well outside their domain, and do not rise to the level of conscious thought. The centipede dilemma seems topical here.
As for one of your examples, autonomous driving, my guess is that computers are better at it than most human drivers in 99% of all cases, but the complexity of the remaining 1% is what kills it.