@bootlegrydia@social.treehouse.systems @pawv@tech.lgbt he is also an AI bro who proposed that job applicants should be evaluated by their llm chat history or somethingI had to look this up. it's even dumber than that. He proposes that "A multi-year chat history is a good indicator", but also, that an perspective employee should submit an LLM that is trained on said chat history. So that the employer's LLM can interview it, about you. This will somehow not reveal private information and have far higher "fidelity"....Which about craters my opinion of John Carmack's intelligence, character aside. A guy who uses LLMs and worked at Meta and at a company attempting AGI has no excuse to not know bad of a take this is.
P
