x
Operator Cognitive Entropy as a Safety Parameter in Human-AI Loops — LessWrong