Ilya Sutskever, co-founder and long-serving chief scientist of OpenAI, has left the company, as announced by OpenAI CEO Sam Altman in a Tuesday evening post on X.
“This is a great loss for us; Ilya is undoubtedly one of the brightest minds of our era, a beacon in our field, and a close friend,” Altman expressed. “OpenAI would not be where it is today without his contributions. While he is embarking on a personally significant endeavor, I remain deeply thankful for all he has achieved here and am committed to completing the mission we began together.”
Jakub Pachocki, OpenAI’s director of research, has been appointed as Sutskever’s successor. Pachocki joined OpenAI in 2017, initially as a research lead for the Dota team, which developed an AI system capable of outperforming human players in Valve’s Dota 2 strategy game. He later became the research lead for OpenAI’s reasoning and science of deep learning organizations before rising to the position of principal research scientist.
It remains uncertain if Pachocki will also assume leadership of OpenAI’s Superalignment team, previously overseen by Sutskever and Jan Leike. According to the New York Times, Leike has also departed from OpenAI.
The Superalignment team was established in July to devise methods for steering, regulating, and governing “superintelligent” AI systems — systems theorized to possess intelligence far superior to that of humans. The Times reports that John Schulman, another OpenAI co-founder, will take over as its supervisor.
TechCrunch has learned that the Superalignment team will be more deeply integrated across OpenAI’s research divisions to enhance its objectives. Consequently, the team as it currently stands may undergo changes in the future.
Greg Brockman, OpenAI’s president, remarked on X that Sutskever “played an instrumental role in building the foundation of what OpenAI has become today.”
Sutskever’s departure follows the rollout of OpenAI’s latest flagship generative AI model, GPT-4, and significant enhancements to the company’s popular AI chatbot, ChatGPT. His exit marks the conclusion of a tumultuous period that began last November.
Just before Thanksgiving, Sutskever and OpenAI CTO Mira Murati voiced their concerns about Altman’s leadership to members of OpenAI’s former board of directors. The issues revolved around strategic disagreements, with Sutskever reportedly frustrated by Altman’s rapid product launches at the expense of safety-focused research.
The old board, of which Sutskever was a member, made the unexpected decision to dismiss Altman without informing most of OpenAI’s employees. They cited Altman’s lack of consistent transparency with the board as the reason for his firing.
This decision enraged Microsoft and other investors, jeopardized OpenAI’s stock sale, and led to a majority of OpenAI’s staff — including Sutskever, in a surprising reversal — threatening to resign unless Altman was reinstated.
Altman was eventually reinstated, and much of the old board stepped down. Sutskever did not return to his role thereafter, as reported by The Times; Pachocki has been acting as the chief scientist since November.
Sutskever, who completed his Ph.D. in computer science under AI pioneer Geoffrey Hinton at the University of Toronto, joined OpenAI in 2015 after leaving Google Brain. He has made significant contributions to AI, including his work on the ImageNet computer vision system and DeepMind’s AlphaGo.
While Sutskever has not yet disclosed his future plans, he conveyed confidence that OpenAI will develop artificial general intelligence (AGI) that is both safe and beneficial. “I am excited for my next venture, a project that holds deep personal significance, which I will reveal in due time,” he stated. “It has been a privilege to work with everyone at OpenAI, and I will miss the team dearly.”
