Meta hires Thinking Machines
- Meta hired five founders from Thinking Machines Lab as part of an aggressive talent push into advanced AI and systems work. - One hire was reported in coverage as tied to a roughly $1.5 billion compensation package. - The raid shows hyperscalers still spend heavily for top-tier AI and systems talent even amid broader workforce restructuring (thenextweb.com).
Meta has hired five founding members of Mira Murati’s Thinking Machines Lab, deepening its push to build advanced artificial intelligence systems and the infrastructure behind them. (thenextweb.com) The hires reported by The Next Web include Andrew Tulloch, Lucas Beyer, Alexander Kirillov, Jiahui Yu, and another founding team member from the startup Murati launched after leaving OpenAI. Coverage tied Tulloch’s move to a compensation package previously reported at roughly $1.5 billion over six years. (thenextweb.com) Thinking Machines says it is building frontier AI systems that are more customizable, multimodal, and collaborative, and says its team helped create products and projects including ChatGPT, PyTorch, OpenAI Gym, Fairseq, and Segment Anything. The startup also says infrastructure reliability and efficiency are core priorities for its research work. (thinkingmachines.ai) Meta is hiring into that effort as it rebuilds its own AI stack. On April 8, Meta said Muse Spark was the first large language model from Meta Superintelligence Labs and said larger models were already in development. (about.fb.com) The scramble for researchers is happening while Meta is also cutting elsewhere. Reuters reported that Meta plans a first wave of layoffs on May 20, 2026, affecting about 10% of its global workforce, or close to 8,000 employees, with more cuts possible later in the year. (tbsnews.net) Thinking Machines is not a small research collective. In March, it announced a multi-year partnership with Nvidia to deploy at least one gigawatt of next-generation Vera Rubin systems for frontier model training, and Nvidia said it had also made a significant investment in the company. (thinkingmachines.ai) The startup has also been shipping product. Its Tinker platform moved to general availability in December 2025 and offers developers tooling to fine-tune open models, including support for reinforcement learning and vision inputs. (thinkingmachines.ai) Meta has not publicly framed the hires in detail on its newsroom, and Thinking Machines has continued to present itself as an active company with products, research posts, and open roles. The immediate result is that Meta added elite AI builders while a rival lab kept operating after losing part of its founding bench. (about.fb.com, thinkingmachines.ai) The dealmaking around a handful of engineers now looks closer to sports free agency than ordinary recruiting. Meta is trimming thousands of jobs, shipping new models, and still paying up for the people it thinks can move the frontier fastest. (tbsnews.net, about.fb.com, thenextweb.com)
