David Robinson, ex-OpenAI safety and policy: SV lacks a safety-centric culture; labs must study other fields' safety approaches; time for trial and error's over (David Robinson/The Atlantic) | TickerVault