Lilian Weng, a co-founder of Thinking Machines Lab and a prominent artificial intelligence researcher, has left the startup after saying its pace and workload had become unsustainable for her health. Her departure was followed almost immediately by confirmation that she would return to OpenAI, where she previously held senior research and safety roles. The rapid transition has drawn attention across the AI sector, highlighting both the intense demands placed on startup founders and the continuing competition among leading laboratories for experienced researchers.
Weng Steps Away From Thinking Machines Lab
Weng announced her departure on July 27, sharing a message originally sent to colleagues at Thinking Machines Lab, which employees informally call “Thinky.” She said persistent stress and heavy workloads had affected her physical health over the previous seven months, leaving her unable to continue operating at the pace required by an early-stage company. Although she did not disclose a specific medical condition, Weng explained that she had considered taking a narrower position inside the company before deciding that leaving the co-founder role was the more appropriate choice.
A Return to OpenAI
On July 29, OpenAI confirmed that Weng would rejoin the company and lead a senior team focused on accelerating internal research. The group is expected to support work across research teams on recursive self-improvement, an area concerned with developing AI systems capable of contributing to the improvement of their own methods and capabilities. Her appointment suggests that Weng intends to remain deeply involved in advanced AI research while moving into a role with clearer responsibilities and less of the broad operational pressure associated with founding a startup.
An Established Research Career
Weng first joined OpenAI in 2018 after earlier positions at Dropbox and financial technology company Affirm, building a reputation through work spanning robotics, applied research, and AI safety. During nearly seven years at OpenAI, she helped establish the Applied AI Research team and later led safety systems work connected to the evaluation and deployment of increasingly capable models. She left the organization in November 2024 and subsequently became one of the co-founders of Thinking Machines Lab, the AI company launched under former OpenAI chief technology officer Mira Murati.
Implications for Thinking Machines Lab
Weng’s exit removes another senior figure from a closely watched startup seeking to compete with established AI developers through new models and research infrastructure. Murati publicly supported Weng’s decision, thanking her for their work together and emphasizing the importance of prioritizing her health. The departure follows broader turnover at Thinking Machines Lab, with Business Insider reporting in May that 13 of its 42 founding members, including three of its original six co-founders, had left the company since its launch.
Weng’s departure and swift return to OpenAI underline the distinction between leaving high-intensity entrepreneurial leadership and withdrawing from the AI industry itself. Her new position allows OpenAI to regain a researcher with extensive institutional knowledge in safety, applied AI, and model development at a time when experienced technical leaders remain highly sought after. For Thinking Machines Lab, the move represents a notable leadership loss, while for the broader sector it offers a visible example of how health, workload, and role design are increasingly shaping careers at the frontier of artificial intelligence.