Warum wurde Sam Altman 2024 gefeuert?
Casey Newton schreibt in Platformer:
I realize that, for most people, this is a settled question. The board fired Sam on November 17, 2023, alleging that he had not been "consistently candid in his communications with the board." Altman and his allies rallied behind him; more than 95 percent of OpenAI employees threatened to quit if he was not reinstated. Three days later, OpenAI brought him back, along with company president Greg Brockman, and all but one member of the board that had fired him resigned.
The board clearly overplayed its hand, struggled to communicate its position, and failed to anticipate the degree to which their move would endanger OpenAI's very survival. Whatever their reasons, they failed to be good stewards of the institution.
But imagine it had gone differently. Imagine that the board had told us in the moment everything they knew at the time, at the same moment they had announced Altman's firing. What would they have said? And what would it have changed?
Knowing what we know now, we know the board might have said: Altman did not inform the board that the company had developed and planned to release ChatGPT. He claimed that three enhancements to GPT-4 had been approved by a joint safety board the company had established with Microsoft; in fact, only one had been approved. Altman also let Microsoft test GPT-4 in India without the safety board's approval.
Imagine if the board had said that Altman also owned the OpenAI Startup Fund, an investment vehicle, personally — something else he hadn't told them about. (He explained that the fund was in his name for tax purposes, and meant to be a temporary arrangement, but the situation had persisted for two years.)
Imagine if the board had said that they were responding to concerns from top executives, who had complained about his leadership for months, and supplemented their allegations with dozens of examples, many of which were supported by screenshots from Slack.
Finally, imagine the board had said that Altman had lied to one board member (Ilya Sutskever) in an effort to push out another board member (Helen Toner), who had been critical of his leadership publicly.
This is the account of events presented this weekend in an excerpt of Keach Hagey's forthcoming book, The Optimist: Sam Altman, OpenAI, and the Race to Invent the Future, in the Wall Street Journal. It's the best account I've yet read of the period leading up and after Altman's firing, and offers the detailed explanation that all of us were craving for months after Altman got his job back. Some details have been previously reported elsewhere, but the collective force of Hagey's reporting is to answer once and for all what it meant when the board said Altman had not been "consistently candid."
As Zvi Mowshowitz put it: "If you lie to board members about other board members in an attempt to gain control over the board, I assert that the board should fire you, pretty much no matter what."