Proposal: All AI agents should be RLHFed based on user responses such as the following: “slop” “cringe” “Reddit” View original