About
Yuan Cao is co-founder and CEO of Unreasonable Labs, a Cambridge, Massachusetts startup building an AI discovery engine that composes new knowledge across chemistry, materials science, biology, and beyond. He was previously a Senior Staff Research Scientist at Google DeepMind, where he made core contributions to the Gemini model family and Google Search's AI Mode, and co-authored the influential reasoning frameworks ReAct and Tree of Thoughts. His earlier career spans Google, Baidu, and the speech-AI startup KITT.AI, following a PhD in Electrical and Computer Engineering from Johns Hopkins University. In 2025 he left DeepMind to co-found Unreasonable Labs with MIT professor Markus Buehler; the company emerged from stealth in March 2026 with a $13.5 million round led by Playground Global to build 'superintelligence for knowledge discovery.' He argues that the real bottlenecks of AI for Science are neither compute nor data but physical-world verification speed and the invention of new concepts, and advocates hybrid LLM-plus-symbolic approaches to push models beyond recombining existing knowledge.
Key Contributions
- Co-founded Unreasonable Labs (2025) with MIT professor Markus Buehler, building an AI discovery engine that composes new knowledge across chemistry, materials science, and biology
- Core contributor to Google's Gemini model family and Search AI Mode as Senior Staff Research Scientist at Google DeepMind
- Co-author of Tree of Thoughts and ReAct, two influential frameworks for LLM reasoning and deliberate problem solving
- Raised a $13.5 million round led by Playground Global (2026) to scale 'superintelligence for knowledge discovery'
- Advocates hybrid LLM-plus-symbolic approaches to AI for Science, arguing the field's real bottlenecks are verification speed and concept invention
Videos & Interviews
Papers & Publications
Tree of Thoughts: Deliberate Problem Solving with Large Language Models
2023Generalizes chain-of-thought into a search over trees of intermediate reasoning steps, letting language models explore, evaluate, and backtrack while solving problems
Read PaperReAct: Synergizing Reasoning and Acting in Language Models
2022Interleaves reasoning traces with actions so language models can plan, consult tools, and adjust — a foundation of today's agent designs
Read PaperConnections
Jeff Dean
CollaboratedCo-founder, Discovery Loop
Cao is one of thirty-one authors of Google's 2016 neural machine translation paper, the system that moved Google Translate onto deep sequence-to-sequence networks, and Jeff Dean's is the last name on it. A decade later, in his 2026 interview, Cao describes Dean's Discovery Loop — letting AI do its own knowledge discovery and automated research — as 'very consistent with the direction our startup set out to pursue.' He also recalls often seeing Dean pair-coding in the office.
arxiv.org · youtube.com
Alan Turing
In contrastMathematician & Computer Science Pioneer
Turing bet that thinking could be mechanized — that whatever a mind does, a suitable machine could in principle do too. Cao, who helped build some of today's most capable reasoning machines, takes the other side at one point: he suspects that abstracting a new concept, the way someone looking at three birds and four trees once invented number, may not be Turing-computable at all. On this view the brain at bottom may not be a Turing machine, and concept invention is AGI's 'last mile', possibly one that can never be crossed.
youtube.com
