The OpenAI Board Fight, Reconstructed

Written days after the November 2023 firing and reinstatement of Sam Altman, Zvi Mowshowitz's piece is still the most careful contemporaneous reconstruction of what happened and why. He frames it not as personality drama but as a structural conflict baked into OpenAI's design: a capped-profit company governed by a nonprofit board with a safety mission. Zvi lays out the specific frictions the board reportedly weighed, including Altman's reaction to a research paper by board member Helen Toner that noted Anthropic had done more credible costly signaling on safety, which led to an attempt to remove her, alongside concerns about lobbying to soften the EU AI Act, aggressive API price cuts, and outside fundraising activity while CEO. He is careful to separate what is known from what is inferred, and argues the resolution, a new board built around Bret Taylor, Larry Summers, and Adam D'Angelo, was not the clean Altman victory it was widely reported as. He reads it as governance structure losing to commercial momentum.

Why it matters

This is the primary-source-grounded account of the most consequential AI governance event so far, and the structural reading still holds up. The conflict it describes, safety mandates versus the money and momentum of scaling, keeps reappearing in new forms, so if you only remember the headline you are missing the part that recurs: mission-bound boards tend to lose to capital.