For a long time, I had the ideal job. As a teacher, I passed on not just mathematics, but also the taste for critical thinking, for confrontation, and the joy of discovery. As a researcher, I contributed my small part to the greater endeavour, alongside the wider mathematical community. I was part of this collective adventure, which I believed would never end. Slowly, insidiously, however, some flaws began to appear. We elevated colleagues to the status of heroes and heroines for their technical prowess and for the number of their publications in so called prestigious journals. We awarded an ever-increasing number of prizes, with ever-larger sums, to a small number of individuals, organisations or institutions, often focusing more on their productivity and reputation than on the substance of their work, or its inherently collective nature. We watched, amused or even flattered, as a Foundation placed a prize in dollars on a small list of mathematical problems to be solved within the millennium. And we used international competition and the lack of institutional funds as excuses to embark on a frantic race for publication, without taking the time to reflect on the fundamental purpose of our work. Over the past few months, doping has become a major issue in our profession. You can obtain it by yourself, or sometimes the institution provides unlimited access to it. At first, it was simply a matter of asking ChatGPT to check a bibliography, rephrase a paragraph, proofread a calculation or improve it. Then it became possible to ask directly to perform the calculation, or even to work out the correct calculation to perform. Now we can act as ‘super-prompters’, putting hundreds or thousands of agents to work to explore as many avenues as possible, having them correct each other’s work, letting them come up with ideas, and thus ‘crack’ one after the other doctoral subjects, problems that colleagues are working on, or those that the Clay Foundation has considered to be the most important. The ecological cost of this new way of practicing
mathematics is not addressed. What price is the mathematical community prepared to make humanity pay to answer its questions? Soon, the AI industries will have grown tired of testing and refining their models on our questions, using our very work we make available to them. In the meantime, we will have turned our profession into an ultra-doped race. In the face of this transformation, it is both urgent and essential that we return to the original purpose of our work and rethink the way we assess it. I dream of a new way of practicing mathematics, one that does not require us to resort to ‘doping’ just to stay in the race. It would be ‘slow maths’, just as there is ‘slow food’ or ‘slow tourism’. A form of maths where we have a firm control of its structure and objectives. Maths that allows us to stray down side paths which sometimes turn out to be more beautiful and more interesting than the one we originally set out to explore. Maths where we take the time to reflect, to write, to explain, and to share. This practice of ‘slow-math’ is essential because it is not the final result that matters and enables us to go further, but rather the process itself, the detours, and the collective endeavour. I am convinced that it is ‘slow-math’ that will help society progress collectively and soundly, and enable young people to continue wanting to learn, to question and to understand. This practice is now only possible if we collectively decide to put an end to the race to publish, and if we agree to value the idea that can lead to a result just as much as the techniques that finalize the proof, and just as much as proofreading, rewriting, and the sharing of ideas and methods. Only then will we be able to leave the race behind, take the time to understand things in depth, and once again practice the ideal profession.
Received 29 September 2026.
Add to the discussion