Ilya Sutzkever has announced of X, formerly known as Twitter, that he is leaving OpenAI nearly a decade after co-founding the company. He is confident that OpenAI “will build [artificial general intelligence] it’s both safe and useful” under the leadership of CEO Sam Altman, President Greg Brockman and CTO Mira Muratti, he continued. In his own post about Sutkever’s departure, Altman called him “one of the greatest minds of our generation” and credited him for his work with the company. Jakub Pachocki, former director of research at OpenAI, who led the development of GPT-4 and OpenAI Five, has took over the role of Sutzkever as Principal Research Associate.

While Sutskever and Altman praised each other in their farewell messages, the two were embroiled in the company’s biggest scandal last year. In November, OpenAI’s board of directors abruptly fired Altman and company president Greg Brockman. “[T]he no longer trusts him [Altman’s] ability to continue leading OpenAI,” the ChatGPT creator announced at the time. Sutskever, who was a board member, was involved in their firing and was the one who asked Altman and Brockman to separate meetings where they were informed they would be fired. According reports emerged at the time, Altman and Sutkever clashed over how quickly OpenAI was developing and commercializing its generative AI technology.

Both Altman and Brockman were reinstated just five days after being fired, and the original board was dissolved and replaced with a new one. Just before that happens, Sutzkever published on X that he “deeply regrets[tted his] involvement in the actions of the board” and that he would do everything possible “to bring the company back together.” He then stepped down from his role as a board member and, while remaining chief scientific officer, New York Times says he never went back to work.

Sutskever said he is moving on to a new project that is “very personally meaningful” to him, although he has yet to share details about it. As for OpenAI, it recently unveiled GPT-4o, which it claims can recognize emotions and can process and generate output in text, audio, and images.