What does GPT-3 tell us about how our brains work?

I just read an interesting article (link below) about how it does on the turing test. I've had this inclination for a while that state of the art AI is "incomplete", in the sense that we have some of the systems to make AGI, but not all of them. One of the comments they make is that "GPT-3 often finds it easier to write code to solve a programming problem, than to solve the problem on one example input", and that's the nail on the head. We can codify situations, describe the rules, put them in memory and run those rules in our head. We can manipulate the input to see how it'll change, we can spot from a problem statement what the rules are instead of focusing on what the answer is. Anyway, light bulb moment shared.

Link: https://lacker.io/ai/2020/...