Yeah, it's strange to me that the default assumption is that current LLMs are already human-level AGI.