Empire of AI: Dreams and Nightmares in Sam Altman's OpenAI

After the release of the New York magazine profile on Sam Altman, OpenAI board member Helen Toner met with Chief Technology Officer Mira Murati. Murati had worked for three years at Tesla before joining OpenAI. She was well respected for her problem-solving abilities and her history of “cleaning up” Altman’s “messes.” He became even more difficult to work with after ChatGPT’s release made him into a public figure. He continued to tell people what they wanted to hear in the moment, sometimes lying or withholding information in the process, then cutting people off when they challenged him. For instance, rather than mediating a dispute between Sutskever and researcher Jakub Pachocki, Altman privately took Pachocki’s side without communicating his feelings to Sutskever. This resulted in each of their teams working on parallel or intersecting projects without clear direction. Murati confronted Altman about these and similar leadership issues. Altman was defensive and iced Murati out, convinced she was reporting these issues to the board. Murati met with Toner in September 2023 to discuss these issues. She also expressed her concerns that OpenAI was shipping products without sufficient testing. Murati then followed up with Toner and encouraged the board to find an independent director who was not an Altman ally.