in

OpenAI’s initial new board counts Larry Summers among its ranks


Meet OpenAI’s new board of directors: Bret Taylor, Larry Summers and Adam D’Angelo. Or, more precisely, the board for the time being.

Around 1 a.m. Eastern Time on Tuesday, OpenAI announced that, after the company’s previous board of directors abruptly fired Sam Altman as CEO last Friday, it had reached an agreement “in principle” for Altman to return to OpenAI as CEO in tow with a new “initial” slate of board members. (He’ll replace interim CEO Emmett Shear, who, along with OpenAI CTO Mira Murati, had the briefest of tenures at the reigns.) Taylor, the former co-CEO of Salesforce, will chair this board, alongside Quora CEO D’Angelo — a holdover from OpenAI’s previous board — and Summers.

“Initial” implies that the board is transitionary rather than permanent. And — this being an “in principle” ordeal — it’s far from concrete. We’ll have to await clarification from OpenAI, which presumably will come at a more reasonable hour in the workday.

If a Verge report is to be believed, however, the final OpenAI board will have nine members total — likely including Altman and a Microsoft exec. Microsoft had reportedly been mulling whether to push for a board seat prior to Tuesday’s breakthrough, which could invite regulatory scrutiny given the company’s relationship with OpenAI. The tech giant evidently feels it’s worth the governance and oversight guarantees.

Now, Taylor’s name had been bandied about in reports over the past several days as a potential appointee to a new OpenAI board. And it’s not out of the question that D’Angelo, who was allegedly deeply involved in negotiations to bring Altman back into the OpenAI fold, has the necessary support to retain his seat. But Summers is a bit of a wildcard — at least on first glance.

An economist, Summers served as the U.S. secretary of the treasury from 1999 to 2001 and as director of the National Economic Council from 2009 to 2010. More recently, he directed the White House U.S. National Economic Council for then-President Barack Obama, where he played a key role in charting the Obama administration’s response to the Great Recession.

Now, you might ask, what’s an economist and political veteran doing on the newly-formed OpenAI board? Well — a tech outsider on the startup’s board isn’t without precedent, first off. Republican member of the House of Representatives Will Hurd held a seat at one point in time, which he relinquished for an ill-fated presidential run.

But Summer’s appointment is also strategic, my colleague Ingrid Lunden pointed out to me via text. With OpenAI increasingly under a policy microscope, Summers brings connections OpenAI will need — and want — with governments, businesses and academia.

Ilya Sutskever, OpenAI’s chief scientist, is one of the big losers here. Reportedly among the board contingent that pushed for Altman’s removal, he’s seemingly been forced to give up great influence at the company he helped to co-found with Altman roughly eight years ago. If his recent post on X (formerly Twitter) is anything to go by, he greatly regrets it. I’d think anyone in his position would.

Tech entrepreneur Tasha McCauley and Helen Toner, director at Georgetown’s Center for Security and Emerging Technologies, are also out. If reporting last night is anything to go by, Altman will be especially pleased by the latter’s ouster; Altman is said to have attempted to have Toner removed from the board earlier this year over a paper she co-authored that cast a critical light on OpenAI’s safety practices.

As for Brockman, who resigned as OpenAI’s president on Friday in protest of the board’s decision to can Altman, it was initially unclear as to what his fate might be. The old board kicked Brockman off, and OpenAI’s announcement Tuesday said nothing of his reappointment.

But Brockman confirmed in a tweet that he’ll be “returning to OpenAI,” albeit perhaps not as its president.



Sam Altman Reaches Agreement to Return to OpenAI

Sam Altman Reaches Agreement to Return to OpenAI

Artificial Intelligence Robot Thinking Desk

Why ChatGPT Isn’t Truly “Intelligent”