An 87-Year-Old Mathematical Riddle
First formally proposed by German mathematician Ott-Heinrich Keller in 1939, the Jacobian Conjecture deals with polynomial maps—systems of equations that transform a set of input numbers into a set of output numbers. In simple terms, the conjecture asked
a fundamental question: if a transformation is locally reversible at every single point, does that guarantee it's also globally reversible? Imagine stretching a rubber sheet. The conjecture suggested that as long as you don't tear or fold it (a condition checked by a mathematical tool called the Jacobian determinant), you should always be able to reverse the stretch and get back to your starting point. It seemed intuitively true, yet for 87 years, it resisted every attempt at proof, becoming infamous for the many flawed arguments published in its name.
Enter Claude Fable 5
The breakthrough came not from a dusty library but from a cutting-edge AI. On July 19, 2026, mathematician Levent Alpöge announced that he, with the help of Anthropic's AI model Claude Fable 5, had found a counterexample. Instead of trying to prove the conjecture true, as generations of mathematicians had, Alpöge and the AI searched for a specific instance where it failed. The result was a surprisingly compact set of polynomial equations, just 216 characters long, that met the conjecture's conditions but was not, in fact, reversible. The AI system helped construct a map in three-dimensional space where three entirely different input points all led to the exact same output. This single, verifiable example was enough to topple the entire universal conjecture.
A Swift and Decisive Verdict
Unlike flawed proofs that can take years to debunk, this counterexample was stunningly easy to verify. Within hours of Alpöge's post on social media, mathematicians around the world were running the numbers through computer algebra systems and confirming the result. The calculation checked out: the map had a constant, non-zero Jacobian determinant, yet it was not globally invertible. The finding immediately disproved the conjecture for three dimensions and, by extension, all higher dimensions. However, the original two-dimensional version of the problem, first posed in 1884, remains an open question. The AI's discovery didn't solve everything, but it decisively closed the book on the general form of the problem.
The New Frontier of AI and Human Collaboration
This event is being hailed as a landmark moment, not just for mathematics, but for the future of scientific discovery itself. While AI models have defeated humans in games and assisted with data analysis, their ability to generate original, creative insights in pure mathematics was less certain. The disproof of the Jacobian Conjecture suggests a new paradigm: a partnership where human intuition guides an AI's powerful, tireless, and unbiased search capabilities. Humans are adept at asking the right questions and recognizing what's important, while AI can explore vast logical spaces without the inherited assumptions that can sometimes hinder human thinking. The process remains somewhat mysterious—Alpöge has not detailed the exact prompts or workflow—but the result speaks for itself.
