> Maybe it's circular. Sure, if not directly circular then consider that our simulations today are used to explore choices/behavior before doing something, so

> Maybe it's circular. Sure, if not directly circular then consider that our simulations today are used to explore choices/behavior before doing something, so it stands to reason that as each universe sim gets to some interesting data, the sim operators learn something, then change their behavior accordingly and tweak the sim and re-run, learning something new in turn, followed by more sim tweaks, ... etc. We today learn new insights from AI and then change our behavior, pour $billions and dev time into designing better AI, etc...so there's definitely a symbiosis relationship. And I expect that we will eventually want to design AI that designs better AI than we can ourselves, which will teach us how to design better AI-designing-AI...etc. :)