I literally just prompted it with “Please finish this essay” and then typed in what justqueenthoughts had posted. And yeah, I think it followed along in the style of the prompt reasonably well, but didn’t really add anything interesting. Which is no surprise, that’s what LLMs do. Point it in a direction and it will supply an approximation of the mean in the training data that applies.

I have to admit, though, I hadn’t heard of the “Two Cultures” debate, popularized by C.P. Snow.