Like many mathematics enthusiasts, I was profoundly shocked—and I am not exaggerating—when OpenAI announced on September 8 that the regularity problem for the Navier–Stokes equations, the fundamental equations of fluid mechanics, had been solved. (The question was whether smooth solutions to these equations remain regular or whether they can instead develop singularities in finite time—in other words, whether the fluid’s velocity or pressure can become infinite.)
In two dimensions, the global regularity of solutions had already been proved by human beings, people like us, our fellow men. In three dimensions, however, the question remained unanswered. It was one of the seven outstanding mathematical problems selected by the Clay Mathematics Institute in 2000, with a prize of one million dollars offered to whoever could solve it.
And now OpenAI, that enemy of mankind, announces that it has proved a fluid can indeed develop a singularity in finite time. Problem solved. But how? OpenAI assembled a system of some 10,000 (!!) AI agents working in parallel. After 88 hours of intensive computation and more than 2.7 million messages exchanged, the proof was complete.
Read also : OpenAI says AI system solved one of mathematics’ biggest unanswered questions
Had some mathematical genius accomplished this feat, champagne corks would have been popping around the world. We had been waiting a century for this moment! The equations themselves were formulated by the Frenchman Navier in 1822 and subsequently developed by the Irishman Stokes in 1845. In 1934, Jean Leray, a graduate of France’s École normale supérieure, proved that generalized solutions existed but could not establish that they would never develop infinite values over time. From then on, the question of their regularity became one of mathematics’ great unsolved problems. And so it remained for a century…
When I was a student in the maths sup’ preparatory class and later at engineering school, our fluid mechanics professors never failed to remind us that the problem remained unsolved. Sometimes they would add: “Perhaps one of you will solve it someday. You’ll win the Fields Medal, and your name will be known throughout the world…” A collective shiver would run through the lecture hall…
«It is a glory of our species that we have been able to prove fundamental mathematical results, which, fittingly enough, bear the names of those who proved them.»
— Fouad Laroui
And now those thick-headed, anti-humanist brutes at OpenAI have stolen a march on the “real” mathematicians, the ones made of flesh and blood! The trauma is universal. After giving a lecture in Nancy on October 8 entitled “The Mathematical Quest for Infinity,” I discovered that what chiefly troubled the audience was this catastrophe: the intrusion of AI into “the queen of the sciences,” as Gauss called mathematics.
The truth is, we hardly know what to do. Why bother training mathematicians any more if AI is going to take over the field and leave human beings not so much as a crumb?
It is a glory of our species that we have been able to prove fundamental mathematical results, which, fittingly enough, bear the names of those who proved them. Pythagoras’ theorem; Euclid’s proof that there are infinitely many prime numbers; Euler’s astonishing identity (the most beautiful equation of them all); Bayes’ formula (which will help you win at cards…); Cantor’s infinity of infinities; the magnificent theorem proved by the mathematician Emmy Noether in 1915 (every symmetry of a physical system corresponds to a conservation law); Gödel’s inconceivable theorem, enough to drive you quite literally mad… The sheer brilliance of the human mind takes your breath away.
And now these AI imbeciles… What a calamity!
The great David Hilbert once declared: “If I were to awaken after sleeping for a thousand years, my first question would be: has the Riemann hypothesis been proved?” Like the Navier–Stokes problem, this conjecture is one of the seven unsolved problems chosen by the Clay Mathematics Institute, each carrying a million-dollar prize. Formulated in 1859 by the German mathematician Bernhard Riemann, it is the Holy Grail of mathematics. Every mathematician who has tackled it has come away defeated, yet all remain fascinated by it. The very attempt to prove it has led to advances in several scientific fields. Terence Tao, perhaps the greatest mathematician alive, has shared his ideas and progress on the problem with the mathematical community, describing it as extraordinarily difficult.
And now, just a few days ago, we learn that OpenAI has published papers on GitHub claiming a theoretical breakthrough on the Riemann hypothesis. This is too much! These people and their wicked algorithms ought to be thrown into a deep pit, with a great boulder rolled over the top. Nothing is sacred to them. Leave Riemann alone! Pure mathematics is just about the last thing mankind has left to prove that we are not soulless machines.
