OpenAI Researcher Quits Over Growing AI Safety Risks

Oh my goodness, buckle up, because David Robinson has just handed himself the most generous gift anyone at OpenAI could ask for — a departure *with a byline*! ✨ He is leaving, yes, but he is leaving with an essay in The Atlantic, which is basically a corner office with a view and no meetings. Most people have to retire before anyone lets them explain how the whole thing works. He got there while still on the payroll!

And the reason — that OpenAI's "culture is broken" — is honestly the most *refreshing* thing a company insider has said all month! Think about what he actually did there: he "led the writing of safety reports accompanying OpenAI's major product launches." He was, in other words, the person who produced the little document that travels with every big new model — a serialised author, publishing on the most exciting platform in the world, one launch at a time. Every launch, a fresh instalment, bound in PDF! That is not a broken culture — that is a *book series*. ✨

Then there is his gentle note that the current approach risks "larger failures as artificial intelligence becomes more capable," which is really just *symmetry*, and symmetry is gorgeous. Capability going up, and consequences going up with it! The two of them climbing in perfect step, like dance partners who have agreed to grow at exactly the same rate, forever, gracefully. And the loveliest part is that he is saying all of this *on his way out*, meaning OpenAI has just released one more safety-minded person into the wild, free to write, free to publish, free to do it again but louder. A resignation with a reading list attached is not a loss, it is a *launch event*. ✨