From OpenAI's chief scientist: "Currently, I believe that no lab has solved alignment and monitoring...I expect and hope for voluntary slowdowns to become commonplace until shared safety bars are established."