x
AI Safety Thursday: Building an economic model of AI automation — LessWrong