OpenAI just pulled off a notable reversal, rehiring three prominent AI researchers who’d left the company barely over a year ago to join a rival startup. Barret Zoph, Luke Metz, and Sam Schoenholz are returning to OpenAI after a stint at Thinking Machines Lab, according to an X post from Fidji Simo, OpenAI’s applications chief.
All three researchers departed OpenAI in late 2024 to help launch Thinking Machines, the high-profile startup founded by former OpenAI Chief Technology Officer Mira Murati. She left OpenAI around the same time, taking some of the company’s technical talent with her.
Zoph held the CTO position at Thinking Machines but will now report to Simo, who oversees ChatGPT. Before his departure, Zoph led post-training at OpenAI—the techniques the company uses to improve AI model performance after initial development wraps up.
The timing of those original departures was rough for OpenAI. The company was dealing with a wave of exits from its research organization, raising questions about internal stability and talent retention. Getting these three back represents something of a win for OpenAI’s efforts to shore up its technical team.
Meanwhile, Thinking Machines isn’t standing still. Murati announced Wednesday on X that Soumith Chintala, a well-known AI researcher hired in November, will step into the CTO role Zoph vacated.
Thinking Machines is positioning itself among the growing crowd of “neolabs”—startups betting they can develop AI models using approaches that major players like OpenAI and Anthropic might have missed or ignored. The company was last valued at $10 billion. By the end of last year, Thinking Machines was reportedly in talks to raise between $4 billion and $5 billion in new funding, which would push its valuation to at least $50 billion.
OpenAI Brings Back Three Key Researchers From Mira Murati’s Thinking Machines Lab
By Cooper
