I agree that coming up with a new concept that suddenly unifies and clarifies a lot of previously disparate knowledge is the essence of this mindset, even when it is not related to mathematics (Scott Alexander tends to do this). In math such conceptual upgrade struck me a few times, like when learning Linear Algebra (what’s a line?), the Stokes theorem (what is integration?), the basics of the Category Theory (what is a morphism?).
It would feel like a huge bit of progress if the AI community came up with some generalized eye-opening redefinition of a common concept or idea. Maybe something like that has already happened, not sure.
I agree that coming up with a new concept that suddenly unifies and clarifies a lot of previously disparate knowledge is the essence of this mindset, even when it is not related to mathematics (Scott Alexander tends to do this). In math such conceptual upgrade struck me a few times, like when learning Linear Algebra (what’s a line?), the Stokes theorem (what is integration?), the basics of the Category Theory (what is a morphism?).
It would feel like a huge bit of progress if the AI community came up with some generalized eye-opening redefinition of a common concept or idea. Maybe something like that has already happened, not sure.