HN seems to think LLMs are hard science, despite all the evidence they're basically cultural artifacts generated by complex models.