“I took this job because I believe OpenAI is one of the most important companies in the world and it was at risk of being destroyed.”
“I was not firing Sam. The board had already done that. I was trying to save the company.”
“Today I got mass formation confirmation that AI is going to be the most transformative technology in my lifetime and possibly in human history.”
“The board did not do this because of a safety disagreement. They did this because of a breakdown in trust.”
“P(doom) of like 5-50% is enough to take very seriously. We should be spending billions on alignment.”
Emmett Shear occupies one of the most unusual positions in AI history: the man who was briefly CEO of OpenAI during its most chaotic period. His three-day tenure (November 19-22, 2023) gave him an insider perspective on the board crisis that few others share, and his public statements provide a window into the event’s internal dynamics.
Shear’s most significant contribution to the public record was his contemporaneous posting during the crisis. While most participants maintained diplomatic silence, Shear offered real-time commentary that helped the public understand the board’s motivations. His insistence that the firing was about a “breakdown in trust” rather than a specific safety disagreement pushed back against both the narrative that Altman was fired for being unsafe and the narrative that the board was simply power-hungry.
Before the OpenAI crisis, Shear was primarily known as the co-founder and longtime CEO of Twitch, giving him operational experience at a platform-scale company. His willingness to step into the interim CEO role reflected both his assessment of OpenAI’s importance and his relationships with board members who trusted him to navigate the situation.
Shear’s views on AI risk are notably calibrated. He publicly discusses “P(doom)” in probabilistic terms, arguing that even a relatively low probability of catastrophic outcomes justifies massive investment in safety research. This probability-weighted approach to AI risk represents a middle ground between dismissal and panic.
Post-crisis, Shear returned to his role as an investor and advisor. His brief time at OpenAI’s helm gave him unique credibility in discussions about AI governance, and he has continued to comment on the structural challenges of governing organizations developing potentially transformative technology.