This Appears to Be Why Sam Altman Actually Got Fired by OpenAI

It's been nearly 18 months since OpenAI Sam Altman was unceremoniously sacked by the firm before being summarily reinstated just a few days later — and a new book claims to have the tea behind the stunning coup.

In an excerpt from her forthcoming book about Altman, Wall Street Journal reporter Keach Hagey revealed, based on dozens of interviews with insiders, that the November 2023 episode jokingly referred to as the "turkey-shoot clusterf*ck" was ultimately caused by power struggles, politicking, and an alleged pattern of untruths.

The root of the debacle seems, per Hagey's excerpt, to have stemmed from a discovery in the summer of 2023, when an unidentified board member learned by chance that OpenAI's so-called "Startup Fund" was allegedly not disbursing funds to intended investors. After months of obfuscation, the board learned that Altman himself owned the fund — a finding, it seems, that began to unravel mounting doubts in the CEO's leadership.

Not long after Altman's ownership of the fund came to light, OpenAI cofounder and ex-chief scientist Ilya Sutskever placed a rare call to fellow board member Helen Toner, a Georgetown AI safety researcher. Instead of being forthright with his thoughts, he suggested Toner talk to former chief technology officer Mira Murati — and those conversations apparently got the ball rolling in earnest.

In a subsequent call, Murati regaled Toner with stories of the CEO's allegedly toxic managerial style. For weeks, the erstwhile CTO said, Altman insisted on bringing in a human resources representative during their one-on-one meetings until she promised not to share certain information with the board.

Beyond the Startup Fund flap, the mercurial CEO was also accused, the book reports, of other major untruths. He once claimed, in a Slack screenshot Murati later shared with the board, that an OpenAI lawyer told him a specific ChatGPT product didn't need any safety review — but when she followed up, that attorney said he never made such an assertion. Altman also seemingly pushed ChatGPT out to the public in the first place without consulting the board ahead of time.

Following that lore drop, Sutskever revealed his own, similar concerns to Toner. He'd been waiting for years, it seems, for the right moment to convince the board to fire his fellow cofounder amid professional tensions and differences in perspective — and the Startup Fund fiasco seemed to be the right time. After speaking in secret to the rest of the board — save for Altman and his right-hand man, OpenAI president Greg Brockman — the former chief scientist and Murati sent hefty PDF files full of related evidence to the sympathetic board members using, as the book notes, Gmail's self-destructing email function.

During a furtive video call on November 16, four of the company's six board members — Sutskever, Toner, former tech founder Tasha McCauley, and Quora CEO Adam D'Angelo — voted to fire Altman and remove Brockman from the board. According to the author's sources, the board members who'd voted on the ouster told Sutskever they were concerned he may have been acting as a loyalty-testing spy.

Ironically, the board's opaque insistence that Altman had been fired because he was not "consistently candid" was true of the governing body that issued it. Neither Murati, who'd been asked to step in as CEO, nor the board informed staff or the press of the behind-the-scenes machinations — and eventually, Altman loyalists began to view the firing, somewhat rightfully, as a coup.

The rest, as they say, is history — though the power play at the root of Altman's failed ouster shows an alarming degree of discord at a company that claims to be radically changing the world.

More on Altman: Sam Altman Whines That When You Become a Billionaire, "Everyone Hates You For Everything"