23h agoE/acc founder @beffjezos proposes satirical "RLHaHaF" training concept to fix humorless AI modelsThe play on RLHF uses human laughter as rewards.SentimentSentimentPos50%Neg50%Positive users cite GPT-5.5, Grok and Claude as already funny in reply to claims LLMs lack humor, while negative users insult the critic as a whining tech bro.6 comments with sentiment. View comments.