70-Page Confidential Letter Alleges First Charge of 'Lying,' Altman Tells Board 'I Can't Change My Personality'

According to 1M AI News monitoring, Pulitzer Prize winner Ronan Farrow and The New Yorker journalist Andrew Marantz have published a lengthy investigative report based on interviews with over 100 insiders, revealing for the first time two core documents: a roughly 70-page confidential memo compiled in the fall of 2023 by former OpenAI Chief Scientist Ilya Sutskever, and over 200 pages of internal notes accumulated by Anthropic CEO Dario Amodei during his time at OpenAI. Both documents have never been disclosed before.

The Sutskever memo includes Slack messages, HR documents, and phone-captured screenshots (allegedly to evade company device monitoring), starting with a list that reads, "Sam exhibits a consistent pattern..." with the first item being "Deception." The memo accuses Altman of distorting facts to executives and the board and deceiving colleagues on security processes. Sutskever was quoted at the time telling another director, "I don't think Sam is the one who should be holding the button."

The Amodei notes, titled "My Experience at OpenAI" (subtitled "Private, Do Not Share"), circulated among Silicon Valley peers but never disclosed, state that "The problem with OpenAI is Sam himself," alleging that Altman personally denied contractual terms while face-to-face with Microsoft to sign a $1 billion investment deal, even after Amodei read out the terms verbatim.

The report also revealed several previously undisclosed facts:

1. The promised independent investigation upon Altman's return did not result in a written report. The law firm WilmerHale (known for leading investigations into Enron and Tyco) responsible for the investigation only orally briefed two new directors, with the decision not to produce a written report partly based on advice from the private lawyers of those two directors. Insiders described the investigation as seeming to be "aimed at limiting transparency," with some current directors considering a potential need for a "redo."

2. The actual compute power allocated to the Alignment team was around 1%-2% of the publicly promised 20%, with most of it assigned to the "oldest, worst-chipped cluster." When journalists requested to interview a researcher working on existential safety, an OpenAI representative replied, "What do you mean by 'existential safety'? That's not a thing."

3. Around 2018, the executive team earnestly discussed an internal initiative dubbed the "National Plan": to have major nations (including China and Russia) bid for AI technology. Then Policy Lead Jack Clark described its aim as "creating a prisoner's dilemma where all countries would have to fund us." The plan was shelved due to multiple employees threatening to resign.

4. Several Microsoft executives express strong dissatisfaction with Altman. One executive stated, "He distorts, twists, renegotiates, and violates agreements," believing that "there is a small but real possibility he will ultimately be remembered like Ponzi scheme perpetrator Bernie Madoff or FTX founder Sam Bankman-Fried."

After being fired, Altman was asked during a call with the board to admit to his deceitful pattern. He repeatedly said, "This is absurd," and then said, "I can't change my nature." One director present interpreted this as, "The meaning of this statement is 'I have a trait of lying to people, and I won't stop.'" Y Combinator's first cohort member and programmer Aaron Swartz, who passed away in 2013, had warned a friend before his death: "You must understand that Sam can never be trusted. He is an antisocial personality and can do anything." The report mentioned that more than one person actively used the term "antisocial personality" in interviews.

In over a dozen conversations with reporters, Altman denied deliberate deception, categorizing the evolving commitments as "well-intentioned adaptation" to a rapidly changing environment, and attributing early criticism to his tendency to "avoid conflict too much." When asked if running an AI company demands a higher level of integrity, he added, "Yes, it does require a higher level of integrity, and I feel the weight of that responsibility every day."