Basically, consider a high-profile scientific problem (one that is getting enough attention from smart humans and their externalized cognitive infrastructure). How good is human Science at converting the information available about the problem into generalizable models of the

→ View original post on X — @fchollet