Sam Altman May Have a Trust Issue

Sam Altman has spent years insisting AI needs unusually principled leadership. The New Yorker asks whether he's the one to provide it—or the biggest risk of all. In a sweeping investigation, Ronan Farrow and Andrew Marantz detail how OpenAI's CEO went from being quietly pushed out by his own board over alleged "lying" and safety shortcuts, to orchestrating a lightning-fast comeback that left his critics sidelined and his power largely intact. Internal memos from co-founder Ilya Sutskever and early safety leader Dario Amodei, reviewed by the magazine, accuse Altman of serial deception, pitting colleagues against one another, and diluting OpenAI's founding safety commitments as the company chased money, speed, and a trillion-dollar valuation.

• "Even people close to Altman find it difficult to know where his 'hope for humanity' ends and his ambition begins," reads the story. "His greatest strength has always been his ability to convince disparate groups that what he wants and what they need are one and the same. He made use of a unique historical juncture, when the public was wary of tech-industry hype and most of the researchers capable of building (artificial general intelligence) were terrified of bringing it into existence. Altman responded with a move that no other pitchman had perfected: he used apocalyptic rhetoric to explain how A.G.I. could destroy us all—and why, therefore, he should be the one to build it. Maybe this was a premeditated masterstroke. Maybe he was fumbling for an advantage. Either way, it worked."

Sources range from current and former OpenAI staffers to senior Microsoft executives who now privately question whether he can be trusted with technology they believe could reshape the global order. For the full reporting, including the internal documents and on-the-record interviews, read the full story.