The News
OpenAI’s chief scientist and co-founder Ilya Sutskever has left the startup just months after taking part in a failed attempt to boot CEO Sam Altman from the company over safety concerns.
“After almost a decade, I have made the decision to leave OpenAI. The company’s trajectory has been nothing short of miraculous,” Sutskever wrote on X. “I am excited for what comes next — a project that is very personally meaningful to me about which I will share details in due time.”
Before co-founding OpenAI with Altman and Elon Musk, Sutskever co-wrote AlphaGo at Google DeepMind, a breakthrough AI program that could successfully win the notoriously difficult board game Go.
Sutskever ran OpenAI’s superalignment team, which is entrusted in developing powerful yet safe AI. Hours after he said he was leaving, Sutskever’s close colleague on the team Jan Leike announced he would depart OpenAI, too.
Sutskever has been replaced by Jakub Pachoki, who was OpenAI’s Director of Research.
SIGNALS
Sustkever’s departure underscores Altman’s control
Sustkever, whose departure came within hours of the release of OpenAI’s most powerful model yet, ChatGPT-4o, played a key role in briefly pushing Altman out of the company over Thanksgiving last year. At the time, Sustekver was on OpenAI’s board of directors, which had broad oversight of the goings on at the company. He helped to orchestrate Altman’s ouster — Sutskever reportedly felt the company was moving too quickly and not prioritizing safety, The Information reported at the time. Ultimately, Altman’s way of doing business won the day. But that failed attempt to keep OpenAI’s tech in check may backfire, The Atlantic noted: “In a terrible irony, Sutskever’s aborted coup may have made it more likely that a large, profit-driven conglomerate develops the first super-dangerous AI.”
Staying at OpenAI may have been ‘untenable’ for Sutskever
Sustkever has not been seen in the OpenAI office since the failed coup against Altman, according to reports, feeding speculation about what the chief scientist knew about the company’s operations — and its safety. Publicly, Sustkever said he regretted trying to oust Altman. But after the episode, OpenAI’s Director of Research, Jakub Pachocki, reportedly took over Sutskever’s responsibilities —now, Pachoki is firmly in the role. While no one is sure yet where Sustkever will land next, The New York Times’ Andrew Ross Sorkin speculated in his newsletter Wednesday Sutskever could go to work with OpenAI co-founder Elon Musk: “Musk has defended Sutskever’s initial effort to oust Altman, and in December he offered his former colleague a job at his xAI start-up.”