The timing was either deeply ironic or painfully revealing.
On the same day OpenAI published a polished set of policy recommendations for how to handle superintelligence—you know, the scenario where AI outsmarts all of humanity—The New Yorker dropped a long investigation into whether CEO Sam Altman is actually trustworthy enough to see any of those promises through.
Reading them side by side feels like watching someone swear a sacred oath while their background check is being faxed to the room.
OpenAI’s proposal is the kind of document that sounds great in a boardroom. It talks about keeping “people first” as AI begins outperforming even assisted humans. It promises transparency, clear-eyed risk monitoring, and vigilance against scenarios like AI evading human control or governments using it to undermine democracy. Without proper safeguards, the company warns, “people will be harmed.” It’s earnest, it’s polished, and it’s exactly what you’d write if you wanted to sound like the responsible adult in the room.
But the New Yorker piece isn’t about the policies. It’s about the person who would have to enforce them.
And the picture it paints isn’t pretty. The article doesn’t rehash old gossip—it digs into a pattern of behavior that makes insiders nervous. Not about the technology, but about whether Altman can be trusted to follow through on the big promises OpenAI keeps making. The kind of promises that, if broken, don’t just hurt a stock price but could affect literally everyone.
Here’s the thing: OpenAI has been saying the right things for years. They’ve positioned themselves as the safety-conscious alternative, the company that takes alignment seriously while others race to ship products. But there’s a growing gap between what they say publicly and what people inside the organization believe is actually happening.
I’ve seen this pattern before in tech. A charismatic founder builds a narrative of responsibility while the internal culture tells a different story. The difference here is that the stakes aren’t just another startup burning through VC money. We’re talking about a company that claims to be building toward artificial general intelligence—and wants us to trust them with the switch.
The New Yorker piece doesn’t accuse Altman of bad intentions. It’s more subtle than that. It questions whether someone with his track record—the pivots, the sudden departures, the shifting stories—can be relied upon when the pressure really hits. And superintelligence, if it ever arrives, will bring pressure like nothing else.
OpenAI’s policy document says the company can be trusted to advocate for a future where superintelligence means a “higher quality of life for all.” That’s a beautiful sentence. But sentences are cheap. Trust is earned through consistency, and consistency is exactly what the New Yorker piece suggests is missing.
I don’t think this is a simple case of good company, bad CEO. The problems are structural. OpenAI’s unusual governance structure, its shift from nonprofit to capped-profit, the revolving door of safety researchers—these aren’t just Altman’s doing. But he’s the face of it, and he’s the one making the promises.
What bothers me most is the disconnect. You can’t release a document about transparency and accountability on the same day a major magazine publishes a story about how people inside your organization don’t trust you. That’s not bad luck. That’s a symptom of a leadership team that either doesn’t see the contradiction or doesn’t care.
Either way, it’s not reassuring.
OpenAI wants us to believe they’re the ones who will handle superintelligence responsibly. But before we hand them that responsibility, maybe we should ask a simpler question: Can we trust the person in charge to tell the truth when it’s inconvenient? Because the policy documents are nice, but they’re only as good as the people who write them—and the people who enforce them.
And right now, the answer to that question seems less certain than OpenAI would like us to think.
Comments (0)
Login Log in to comment.
Be the first to comment!