A New Kind of Research Partner
World-renowned mathematician and Fields Medalist Terence Tao has offered a fascinating glimpse into the future of scientific research, sharing a public conversation where he uses ChatGPT to attack the famous Jacobian Conjecture. The log, shared on Hacker News, reveals Tao using the AI model not to find a solution, but to act as a highly capable assistant for exploring a potential counterexample. Tao directs the inquiry, feeding the model complex algebraic expressions and high-level mathematical concepts, while the AI performs laborious calculations and symbolic manipulations.
The exchange demonstrates a powerful new workflow for top-tier academics. Tao described the AI's role as that of a "highly proficient, tireless, and fast" research assistant, capable of handling the tedious but critical steps that underpin creative mathematical work. This human-AI collaboration allows the expert to focus on strategy and insight while offloading computational burdens.
The AI's Role: Assistant, Not Author
It is crucial to understand that the AI did not generate the novel mathematical ideas. Professor Tao provided the intellectual framework, the strategic direction, and the core hypotheses. The AI's function was to execute specific, well-defined tasks that are prone to human error and time-consuming to perform manually. This highlights the current role of LLMs in advanced research as powerful tools for augmentation, not replacement.
Key tasks performed by ChatGPT in the session included:
- Performing complex symbolic algebraic expansions.
- Verifying multi-step calculations to ensure accuracy.
- Identifying potential inconsistencies in the mathematical logic.
- Translating abstract concepts into formal notation like LaTeX.
This division of labor—human for creativity, AI for execution—is becoming a defining feature of modern research. For more insights into how AI is transforming science and industry, subscribe to the AI Breaking Wire newsletter. Join over 100,000 professionals who receive weekly updates on the tools and trends shaping our world.
Beyond Code: AI Tackles Abstract Mathematics
While LLMs have proven remarkably effective at generating and debugging code, this example showcases their growing proficiency in more abstract and symbolic domains. The ability to correctly parse and manipulate advanced mathematical concepts, which lack the rigid syntax of programming languages, represents a significant step forward for AI capabilities. The model appeared to follow Tao's reasoning, correctly interpreting nuanced mathematical instructions and contributing meaningfully to the investigative process.
This application pushes the boundaries of what many considered possible for current-generation language models. It suggests that their utility extends far beyond text summarization or software development into the very heart of abstract scientific inquiry, a domain once thought to be exclusively human.