The Nature Noro Science Paper Is Really About Metal Learning

i think the argument is that optimizing a reward on a reward function through through experience with the world, it is guaranteed to be enough. What it doesn't guarantee, and which is something that's neglected in their argument, at least as they made it verbally in inside t mine, right? Like, yes, reward r l is enough to get any intelligent behavior a given infinite experience, infinite data. A, but it doesn't explain how, a, you know, how humans and other animals, for example, are able to learn in such a simple, efficient way.