Post #4149685

(2) has a big “if” in there Its a super big if, but it is one lesswrong (and Anthropic, with their “model welfare” pseudoscience) is allegedly seriously considering as a possibility (since, you know, they think LLMs are already AGI and Eliezer was accusing AI-dungeon, you know, GPT-2, of deliberately scheming). If anything, it speaks to some possible hypocrisy/motivated reasoning that they consider LLMs AGI but not possibly morally relevant. conceptualized capitalism This seems to be a lesson most lesswrongers have had rubbed in their faces repeatedly over the past 5 years with everything about LLM company’s behavior, but they just don’t want to learn.