x
Agent-Structure Problem — LessWrong