OpenAI safety writer resigns, warns culture 'broken' — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- David Robinson resigned from OpenAI this week after writing the safety reports that accompanied every major model release, and is now sounding the alarm in an editorial published by The Atlantic
- Robinson says the AI industry's culture is "fundamentally broken" — a problem deeper than new rules or regulations — blaming Silicon Valley's "extreme confidence," "perpetual sprints," and "unimpeded optimism" about potential harms
- Robinson argues frontier labs should run like nuclear power plants or busy airports, with layers of redundancy and careful, time-consuming planning so that inevitable human error does not open a door to disaster
- Jacob Coxon appears to have kicked off the recent wave of safety departures by quitting Anthropic and publicly warning that AI "could kill us all by the end of the decade"
- Robert O'Callahan, Bilal Chughtai, and Josh Engels subsequently left Google DeepMind, and Joe Benton also departed Anthropic, according to the article
Why it matters: Robinson is the latest in a string of safety researchers to leave frontier AI labs — joining five others from OpenAI, Anthropic, and Google DeepMind in recent weeks — which means the companies racing to build the most powerful models are simultaneously shedding the personnel specifically tasked with flagging risks, exactly when Robinson says those labs need nuclear-industry-level caution instead of Silicon Valley's "perpetual sprints."
Ask SkimNews
