Comment by edot

Has nothing to do with knowing the answer, and everything to do with infinite persistence. I believe any current-generation LLM can, given an environment where it can experiment rapidly, do anything.