i don’t consult LLMs for ‘life advice’. feels kind of cheating. that the process of learning is the fun. but I do consult it for health and fitness advice. and I’ve noticed the models are now at a stage where they’re correct almost entirely. the best answers most of the time.

this makes me conjecture that in the future, people will consult AIs for every decisions. and they will give the best advice. so for example, 2 people meeting for a negotiation will both have had consulted the AIs for their respective best strategies. it’ll effectively be two intelligent LLMs negotiating through human bodies. and since both LLMs are perfect, the negotiation will be game theory optimal. everyone, everywhere will be playing their best strategies.

the question then becomes whether if everyone plays their best strategies, does it lead to the best societal payoffs?