Meta’s loss is Thinking Machine’s gain

Weiyao Wang spent eight years at Meta – his first job out of college – helping build multimodal visualization systems and supporting open source projects, including SAM3D. His last day at Meta was last week, and he joined the Thinking Machines Lab (TML).

Its move to TML comes as the AI ​​startup expands on several fronts. It just signed a billions of dollars and Google, giving them access to Nvidia’s latest GB300 chips and making it one of the first to use hardware.

The agreement, which was announced last Tuesday at Google Cloud Next, follows a previous agreement with Nvidia, placing TML in the same architecture category as Anthropic and Meta. (Meta is said to have discussed finding a Thinking Machine at this time last year and lately he has been removing TML founders one by one.)

The talent picture remains fluid. Wang and Kenneth Li — a Harvard PhD who spent 10 months at Meta before joining TML this month — are the latest examples of talent poaching going both ways. Business Insider reported last week that Meta is now hacked Seven of the founding members of TML. A review of recent hires shows the Thinking Machines attacking the Meta itself. At least, it seems based on LinkedIn profile analysis, that TML has hired more researchers from Meta than other employers.

The most notable is Soumith Chintala, CTO of TML, who spent 11 years at Meta and founded PyTorch, an open source deep learning solution that is now supporting many AI research projects around the world. He left Meta at the end of 2025 and was appointed CTO earlier this year. Piotr Dollár, an 11-year Meta veteran who previously worked as a research director and co-wrote the Segment Anything model, is now on TML’s technical services. Andrea Madotto, a research scientist in Meta’s FAIR division who focuses on multilingualism, joined TML in December. James Sun, a software developer with almost nine years at Meta working on LLM pre- and post-training, also jumped in.

TML has brought out talent from beyond Meta, too. Neal Wu – a three-time gold medalist at the International Olympiad in Informatics and a founding member of buzzy coding startup Cognition – joined earlier this year. Jeffrey Tao came through Waymo, Windsurf, and OpenAI. Muhammad Maaz previously participated in research at Anthropic. Erik Wijmans arrived from Apple. Liliang Ren spent two and a half years on Microsoft’s AI Superintelligence team retraining OpenAI versions of the code before joining in March.

The number of founders has now reached about 140.

The Meta charge – seven figures, wirelessly connected – is well known here. For researchers weighing their options, the calculation can be as simple as this: Thinking Machines Lab is the most important. $12 billion. Although this figure would not have been impossible to imagine by the company at this time in the previous technology trends (he has released recently. single product so far), compared to breaking OpenAI and Anthropic records, there is still a lot of economic potential.

Reached Friday morning, a TML spokesperson declined to comment on the matter.