LLMs are like stupid savants who know a lot of things. Leaving the requisite “no, that’s not what language models are, you’re misunderstanding what’s important

LLMs are like stupid savants who know a lot of things. Leaving the requisite “no, that’s not what language models are, you’re misunderstanding what’s important here, the best knowledge model already exists and it’s called Wikipedia”