OpenAI Is Rapidly Losing Senior Members to Rival Anthropic
In a series of seemingly unrelated departures, OpenAI is losing three of its top leaders. Greg Brockman, the A.I. company’s president; John Schulman, co-head of its post-training team; and Peter Deng, an OpenAI product head, have all recently stepped away from the company in one form or another.
Brockman will be taking time off through the end of the year, he announced in an X post earlier this week. The sabbatical will be his “first time to relax” since co-founding OpenAI some nine years ago, he said. Deng, meanwhile, reportedly decided to leave his position as vice president of product earlier this year. He joined OpenAI in 2023 and formerly held product leader positions at companies like Meta (META) and Uber (UBER).
Schulman, another OpenAI co-founder, has been more transparent about his plans going forward. The research scientist, who recently took a leadership role in OpenAI’s alignment science efforts and was appointed to its new safety committee, is leaving to join Anthropic, one of OpenAI’s emerging rivals.
“This choice stems from my desire to deepen my focus on A.I. alignment, and to start a new chapter of my career where I can return to hands-on technical work,” said Schulman on X. He described Anthropic as a place to “do research alongside people deeply engaged with the topics I’m interested in” and clarified that he is not leaving OpenAI due to a lack of support for alignment research, which focuses on preventing A.I.-enabled societal harm. In a responding X post, OpenAI CEO Sam Altman praised Schulman and called him a “brilliant researcher” and “great friend.”
Anthropic was founded by siblings Dario and Daniela Amodei, both former OpenAI employees who left in 2020 to launch their own company with an emphasis on developing safe and responsible A.I. Anthropic counts Amazon (AMZN), Google (GOOGL) and Menlo Ventures among its investors and is currently valued at $15 billion.
At least 4 OpenAI employees recently defected to Anthropic
Schulman isn’t the only OpenAI employee who has defected from the company in recent months. Two other OpenAI co-founders, Andrej Karpathy and Ilya Sutskever, both left earlier this year and have since launched their respective A.I. startups.
Jan Leike, who formerly co-led OpenAI’s Superalignment team with Sutskever, stepped away from OpenAI in May, also to join Anthropic. At the time, he described his former employer as having prioritized “shiny products” over “safety culture and processes.” The researcher appeared to support Schulman’s recent decision to join Anthropic, writing that he is “excited to be working together again” in an X post.
Anthropic has also recently brought on Pavel Izmailov, a former OpenAI researcher focused on reasoning who worked with the company’s safety team. Izmailov was reportedly fired by OpenAI in April alongside another coworker for allegedly leaking information and joined Anthropic the following month, according to his LinkedIn page. And Steven Bills, a member of OpenAI’s technical staff for more than two years, updated his LinkedIn last month to reveal that he’d left OpenAI to join Anthropic’s alignment team.
Anthropic leadership made an indirect jab at OpenAI’s exodus of talent while speaking at the Bloomberg Technology Summit in May. “We have seven co-founders. Three and a half years later, all of us are still at the company,” said Dario Amodei during the talk. OpenAI’s 11-person founding team, meanwhile, has dwindled down considerably since the company’s 2015 launch. With the loss of Schulman and Brockman, it currently only has two active founding members left: CEO Altman and Wojciech Zaremba, who heads language and code generation teams.