It also completely ignores what we know about evolution. Our brains are the result of billions of years of natural selection. The amount of “training data” that resulted in our neural structures is on a completely different scale than the training data used for today’s LLMs. And this isn’t even up for debate.

The assumption that this process can be “distilled” from written word is completely insane. I’m not sure how people trick themselves into thinking it’s even remotely possible.