Assuming most people can detect LLM writing quickly. I don’t think that’s true. In this very submission we see people referencing cases where colleagues couldn’t detect something is written by LLM even after reading everything.