Meta’s loss is Thinking Machines’ gain

Weiyao Wang, who dedicated eight years of his career to Meta, has recently made a significant career shift to Thinking Machines Lab (TML). During his tenure at Meta, Wang played a pivotal role in developing multimodal perception systems and contributed to projects like SAM3D. His departure last week coincides with TML's ambitious expansion plans. This move comes at a time when TML is making headlines for signing a groundbreaking multibillion-dollar cloud agreement with Google. This deal, announced during Google Cloud Next, grants TML access to Nvidia’s cutting-edge GB300 chips, positioning the startup among the elite in the AI sector, alongside well-known players like Anthropic and Meta. Interestingly, reports suggest that Meta had considered acquiring TML last year, but the startup has recently seen a surge in talent leaving Meta for its ranks. The ongoing talent dynamics reveal a competitive landscape, with Wang and Kenneth Li — another recent addition from Meta who holds a Harvard PhD — exemplifying the back-and-forth recruitment trend. In a notable twist, Business Insider reported last week that Meta has successfully recruited seven founding members from TML. However, a closer look at recent hires indicates that TML is attracting more researchers from Meta than any other single organization. Among the notable recruits is Soumith Chintala, TML's newly appointed CTO, who spent 11 years at Meta and co-founded PyTorch, a widely used open-source deep learning framework. His departure from Meta in late 2025 was a significant loss for the company. Additionally, Piotr Dollár, another 11-year veteran at Meta, has joined TML’s technical team, bringing his expertise from co-authoring the influential Segment Anything model. The wave of talent migration doesn't stop at Meta. TML has also welcomed experts from various tech backgrounds, including Neal Wu, a three-time gold medalist in the International Olympiad in Informatics, and Jeffrey Tao, who has experience with Waymo, Windsurf, and OpenAI. The startup's current workforce is approximately 140 individuals strong. With Meta’s compensation packages well-known for their generous seven-figure salaries, the decision-making process for researchers appears clear. TML, currently valued at an impressive $12 billion, presents a compelling opportunity for those evaluating their career paths. This valuation, while extraordinary for a startup at this stage, still pales in comparison to the soaring valuations of established names like OpenAI and Anthropic. As the tech landscape evolves, the rivalry between these two entities is likely to intensify. A spokesperson for Thinking Machines Lab declined to provide further comments when approached for this story.