The New Yorker just published an 18-month investigation by Ronan Farrow and Andrew Marantz that reveals what was really happening inside OpenAI before Sam Altman's brief ouster in November 2023. Chief scientist Ilya Sutskever compiled roughly 70 pages of Slack messages and HR documents, sent them to board members as disappearing messages, and one memo opens with a list titled "Sam exhibits a consistent pattern of..." where the first item is "Lying."
The allegations center on how Altman represented OpenAI's safety testing to the board. According to documents reviewed by Farrow and Marantz, Altman reportedly misrepresented deployment readiness benchmarks and downplayed concerning behaviors found during Red Teaming exercises. Sutskever, who once officiated Greg Brockman's wedding with a robotic hand as ring bearer, grew increasingly worried as the company approached human-level AI. His view, as he told another board member: "I don't think Sam is the guy who should have his finger on the button."
Board members Helen Toner and Tasha McCauley saw the memos as confirmation of what they'd already suspected. They fired Altman by video call while he was at a Formula 1 race in Las Vegas. Microsoft, which had sunk $13 billion into OpenAI, found out minutes before the announcement, highlighting the internal dysfunction described by former Azure engineers. CEO Satya Nadella said he was "very stunned." Board member Reid Hoffman told investigators he was "looking for embezzlement, or sexual harassment, and I just found nothing."
Within days, Altman was back. The employee revolt was overwhelming, investors threatened to pull a pending $86 billion deal, and Altman's allies ran what he called a "government-in-exile" from his $27 million San Francisco mansion. This chaos reflects a shift in investor sentiment. Altman is back in charge. The board members who fired him are gone. And OpenAI kept building toward AGI with a CEO that his own chief scientist said couldn't be trusted with the button.