Lilian Weng rejoins OpenAI after leaving Thinking Machines citing health concerns

Jul 29, 2026

AI researcher Lilian Weng is rejoining OpenAI to lead a team on "recursive self-improvement" just days after leaving Thinking Machines Lab, the startup she co-founded. Weng cited health concerns, stating the startup's pace and workload were unsustainable. Her departure highlights an exodus from Thinking Machines, where only two of six original co-founders remain.

Weng's return to OpenAI

▪At OpenAI, Weng will lead a team focused on accelerating internal research to achieve "recursive self-improvement."
▪Weng previously worked at OpenAI from 2018 to late 2024, where she helped develop the GPT-4 model and led Safety Systems work.
▪Lilian Weng, a co-founder of Thinking Machines Lab, is rejoining OpenAI.

Departure from Thinking Machines

▪Weng announced she was leaving Thinking Machines Lab during the week of July 27, 2026, citing health concerns.
▪Weng stated the "consistent stress and workload" at the startup had pushed her "beyond what my health can sustain physically."
▪Mira Murati, CEO of Thinking Machines, publicly supported Weng's decision to prioritize her health.
▪Weng said a "scoped role in a more predictable place, and not as a cofounder" would be a better fit for her.

Recursive self-improvement focus

▪Recursive self-improvement is a process where AI models can train and improve their successors.
▪At OpenAI, Weng will lead a team focused on accelerating internal research to achieve "recursive self-improvement."

Thinking Machines leadership exodus

▪Of the six original co-founders of Thinking Machines Lab, only CEO Mira Murati and chief scientist John Schulman remain.
▪Former co-founders Barret Zoph and Luke Metz left for OpenAI, while Andrew Tulloch departed for Meta.
▪Lilian Weng is one of several co-founders to leave Thinking Machines Lab since its launch.