Source: Meta AI researcher Andrew Tulloch, who in October 2025 came over from Thinking Machines Lab as one of the highest paid employees, is leaving the company
Meta AI researcher Andrew Tulloch, who gained attention for being one of the highest paid employees in tech industry, is leaving the company …
Context & Ripple Effects
The reported departure would unwind a high-profile October 2025 recruitment from Thinking Machines Lab, following Tulloch’s move from the startup to Meta. Meta had also reportedly added two other Thinking Machines Lab founders in February 2026, making the startup a notable source of its AI-research hires.
It also fits a broader retention problem in Meta’s AI organization: several Superintelligence Labs hires reportedly returned to OpenAI in 2025, while OpenAI recruited former Meta researcher Ruoming Pang in February 2026. The repeated moves make talent continuity, rather than the headline value of an individual offer, the salient issue.
First-order effects
- If confirmed, Tulloch’s exit removes a prominent researcher from Meta AI less than a year after his reported recruitment from Thinking Machines Lab.
- Meta’s AI leadership must absorb another reported senior-talent departure alongside its earlier recruiting push from Thinking Machines Lab.
Second-order effects
- The reported exit increases pressure on Meta to demonstrate that its research environment can retain elite hires, after several short-lived Superintelligence Labs appointments and OpenAI’s recruitment of Ruoming Pang.
- Thinking Machines Lab’s founders become a more consequential talent pool for competing AI labs, since Meta had reportedly recruited Tulloch as well as Christian Gibson and Noah Shpak.
Third-order effects
- If rapid moves among senior researchers persist, AI labs will compete less on one-time compensation packages and more on the autonomy, research direction, and organizational stability offered after hiring.
- The pattern points toward a talent market in which individual researchers can repeatedly reprice their value across rival frontier-lab teams, increasing the difficulty of building durable research groups.
The trend: Frontier AI labs are discovering that exceptional compensation can win recruits but does not by itself create durable research-team retention.