A member of OpenAI's product team wrote safety standards will need to be much
higher in the era of intelligent agents, that the safety industry must speed up
to meet that requirement, and that they were glad 'we' are slowing development
to prepare for risks.