Prof. @VUBrussel • From Math to Machines & Back • Writer @tijd @eos_wetenschap & Books • EMS Education • Scientific Council @BelgiumDefence @BELSPO @imec_int
Belgium
Joined November 2009
- Tweets6.8K
- Following339
- Followers2.9K
- Likes17.9K
I’m excited to announce that I will be working at the Mathematical AI Safety Institute (MAISI) full-time as its founding Scientific Director! This means I will not be joining OpenAI as an employee, though I will still be collaborating with them, and I’m happy to announce that OpenAI will be supporting MAISI. They have offered us $10MM in API credits, and access to their latest models for our research use. Moreover, they will be participating in a multi-lab AI safety workshop we are organizing for next year with other frontier AI labs.
As an independent research institute, we are also in conversations with other labs and AI safety orgs about how MAISI can collaborate with them and contribute quickly to AI safety research across the industry. As we’ve seen over the past few weeks, the need for action on AI safety is extremely urgent. Research is an important part of that, and mathematicians have an important role to play!
‘Dit is het perfecte moment om wiskundige te worden: the world needs you’
tijd.be/opinie/column/wiskun…
Ann Dooms retweeted
I’m excited to announce that I will be working at the Mathematical AI Safety Institute (MAISI) full-time as its founding Scientific Director! This means I will not be joining OpenAI as an employee, though I will still be collaborating with them, and I’m happy to announce that OpenAI will be supporting MAISI. They have offered us $10MM in API credits, and access to their latest models for our research use. Moreover, they will be participating in a multi-lab AI safety workshop we are organizing for next year with other frontier AI labs.
As an independent research institute, we are also in conversations with other labs and AI safety orgs about how MAISI can collaborate with them and contribute quickly to AI safety research across the industry. As we’ve seen over the past few weeks, the need for action on AI safety is extremely urgent. Research is an important part of that, and mathematicians have an important role to play!
Peter Hinssen | Liever een raket dan een brandend bushokje tijd.be/opinie/column/peter-…
On my way ✈️
🔢 #BMD2026 starts today!
For the next three days, Barcelona will bring together the mathematical community to share research, knowledge and new ideas.
📅 7–9 October.
📍 Institut d’Estudis Catalans · Barcelona.
Check out the full programme: scm.iec.cat/congres/bmd-2026…
#SCM
If guided wisely, AI will not diminish the mathematician’s role but will broaden it: mathematicians will become designers, curators & interpreters of intelligent
systems that, in turn, help humanity deepen its understanding of the
mathematical universe.
zenodo.org/records/18205480
Ann Dooms retweeted
Bedankt @Zu_Demir om het belang van een bekwaamheidsbewijs te erkennen in de commissie. Het is wel onduidelijk of dit gaat om een vereist of voldoende geacht bekwaamheidsbewijs.
Not to be missed!
Why should mathematicians care about European PISA results?
Join Professor Nuno Crato (University of Lisbon) tomorrow, 9 October, at 16:00 CEST for a lecture and discussion.
Organised by @soccatmat & @euromathsoc Education Committees.
Join via Zoom: us02web.zoom.us/j/8759268063…
Ann Dooms retweeted
Absolutely! This is the PISA paradox: designed to assess competences, it reveals that training students in applied problem-solving falls short, while building deep knowledge leads to good results
“In short, PISA is a screen-based test of problem-solving skills that requires limited specific curriculum knowledge. If you wanted to design a test that would be *least likely* to reward England’s educational approach, and *most likely* to reward the approach in Scotland & Wales, it would be PISA.
And yet, England are still doing better on it.”
🎯 🎯🎯
Ann Dooms retweeted
We’re releasing a broad range of new mathematical results produced by an internal frontier model.
We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.
github.com/openai/math
Ann Dooms retweeted
Major announcement in mathematical formalization!
Finally, after 3 months of very intensive, nonstop work by several AI agents (Codex and Claude Code), we have settled a classification of ALL 15,973 semigroups of order 6. Every one of the 15,969 that has a finite basis now has it written down explicitly, and the remaining 4 are proved to have none at all.
The whole list of bases is produced for the first time in history. Until now, for most of these semigroups, mathematicians only knew that a basis exists; nobody had ever written one down. The formalization in Lean, orchestrated with a multi-agent approach, reached more than 5 MILLION lines of verified Lean code, one of the largest auto-formalization projects to date.
It's been an amazing joint work with @JanotaMikolas and @Jan_Hula, accompanied by senior experts in semigroup theory, João Araújo and Edmond W. H. Lee.
Our paper describing this project, "Proving at Scale for Universal Algebra", has been accepted at the MATH-AI workshop, NeurIPS 2026!
When we launched this project, we were a bit pessimistic: producing proofs for 15,973 semigroups seemed out of scope for the current technology. But with a careful setup (agents communicating through a mailbox we arranged, orchestrating the effort with a custom method of bootstrapping and auto-research), it has finally come to the finish line. The last 390 semigroups took 23 more days. The very last one needed a whole structural analysis before its proof could even be written.
Ann Dooms retweeted
The instructional theory piece is important. Explicit instruction w/o well-designed content lessons won't necessarily be effective.
Ann Dooms retweeted
De besparing op fundamenteel onderzoek, de enige overheidssector die internationaal uitblinkt, de motor van economische groei, de enige structurele oplossing voor onze begroting. Dié besparing is ongeveer (iets meer) wat men uitdeelt aan frigobons. De helft van nutteloze brooddozen. Daar kan men veel lelijke woorden voor bedenken.
tijd.be/opinie/algemeen/univ…
Spot on!
The reason why you have lists of open problems to train LLMs on is because mathematicians have done an excellent job at fostering openness in their professional community, with the goal of advancing mathematics through a combination of collaborative efforts and individual recognition.
The fact that today anyone can prompt a mathematical question (whose formulation may have taken decades to be put in focus by several mathematicians) to get a software-produced, Lean-verified proof (that then needs human work to become intelligible and lead, hopefully, to formulate new problems) is a textbook example of a Pyrrhic victory.
People have already stopped talking about open problems related to their latest research progress when teaching graduate courses or giving lectures because anyone in the room can try their luck at the slot machines of LLM. The intellectual damage already done is vast and deep, and it will take a very deliberate effort to mend it.
As things stand today, you have generated some hundreds of correct proofs (that no human can even give a talk on, because no human has yet understood them), at the price of shutting down the conversation that has provided you with your precious lists of open problems, and more importantly, has generated uncountable progress, including the mathematical foundations of LLM ls themselves: which list of problems were Markov or Shannon trying to solve when they fid their foundational work? And did Riemann or Gauss or Newton had a list of open problems? Did Lagrange invent the Calculus of Variations to solve problem 42 in someone’s list?
How on earth is this considered progress? Looks more like fishing salmons with nuclear bombs until they go extinct. You’ll eventually miss the sushi.
Ann Dooms retweeted
The advisory group on mathematics and artificial intelligence has just published a set of recommendations concerning the responsible release of mathematical results generated by AI companies using internal models. 1/3
agmai.org/general-sep29/
Ann Dooms retweeted
Since all sorts of people have asked me, I decided to write down my current thoughts on "What's the Future for Pure Math Research in the Age of AI?"
writings.stephenwolfram.com/…
