Based on the links posted by c5 above, one purpose of AGI is to be self-contained and self-perpetuating, without needing human interaction. That's both exciting and scary (to me) at the same time. It may also be what people were freaking out about going on behind the scenes with OpenAI's Altman, and why people like Elon, who have warned about potential dangers of AI, were pressing them to divulge further details.