so you're limiting the scope of potential solutions to theorems and then trying to eventually supplant a brute force method with biologically informed cognitive

so you're limiting the scope of potential solutions to theorems and then trying to eventually supplant a brute force method with biologically informed cognitive methods? neural networks? sort of spamming number theory at theorems? by discovering a corpus' bias?