> I would argue that LLMs are actually smarter than the majority of humans right now

This (surprisingly common) view belies a wild misunderstanding of how LLMs work.