#NPnerd fishing in the #ComputationalSea
Uni. Padova, Italy
Joined December 2011
- Tweets4.6K
- Following119
- Followers1.4K
- Likes10.2K
Pinned Tweet
#backpropagation - for poeple suffering from derivative allergy (hopefully useful for teaching @jeremyphoward)
Matteo Fischetti retweeted
1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.
🤖 Made with AI
Matteo Fischetti retweeted
Excited to share that Mercury 2 is one of @TIME's Best Inventions of 2026.
Not long ago, diffusion LLMs were just a research idea. Now Mercury powers products people use every day.
Proud of the @_inception_ai team. Something bigger coming next week 👇🧵
A short video about our new EndoWake model for wind farm layout optimization #orms joint work with @MartiFischetti
Can wind wakes be described by a simple linear convolution?
Presenting *EndoWake*, aka Wind Modelling by Linear Programming; joint work with @MartiFischetti
1. Theory and MILP model:
researchgate.net/publication…
2. Application on GPUs:
researchgate.net/publication…
A digital twin of Fermat in 1637: reconstructing the marvellous proof and locating where it fails
claude.ai/artifact/UMa7Uh3U4…
For all FP-riends #orms:
Feasibility Pump + objective cutoff as a black box: what does it lose? On MIPLIB 2017 inside SCIP, a simple recovery beats SCIP's pump on 116 of 176 instances, loses on 9 (post hoc).
Code and data: github.com/mfischetti-arch/f…
paper: researchgate.net/publication…
Matteo Fischetti retweeted
Over the past few days, I've taken the time to summarize my thoughts on the recent incidents involving agents’ misaligned behavior. We don't know with certainty what comes next, but we know where these issues originate, and this can help us plan the path forward.
Please feel free to ask your questions in the replies, and I’ll try to answer some of them in the coming weeks.
yoshuabengio.org/en/publicat…
Matteo Fischetti retweeted
This could be the most significant problem we are called to address together, by contributing in whatever way each of us can, while upholding integrity and ethics as hard, non-negotiable constraints.
Full declaration and how to endorse it: mathandai.org
Matteo Fischetti retweeted
🏅 Las Medallas SEIO 2025 reconocen a Antonio Cuevas González y a Dolores Romero Morales por sus destacadas trayectorias y contribuciones a la Estadística y la Investigación Operativa. Hoy, ambos medallistas ofrecerán sus conferencias en la #SEIO2026.
¡Enhorabuena! 👏
I found it helpful to have Claude/chatGPT use colored markdown in the chat to highlight positive, negative, and neutral outcomes.
Legenda: ✅ buono · ❌ negativo · ⚠️ da verificare o numero fragile · 🔵 in corso
Matteo Fischetti retweeted
Very excited that we are bringing cheaper / free Claude to more scientists!
If you are using Claude for scientific research, I'd love to hear your feedback, what works well and what doesn't - accelerating science is the main reason I care about AI!
Starting today, 10,000 scientists across every field, from math to chemistry to physics and more, can get Claude through our new Claude Team plan for scientists. Standard seats are free, and premium seats with 5x usage limits are $15 per month, an 80% discount, for one year.
Claude is becoming increasingly capable of scientific work, with recent progress on problems from advanced physics calculations to protein design. Alongside that progress, we've been investing in the research community: Claude Science launched in June, and our AI for Science program funds high-impact projects with free credits. Today's expansion builds on both.
Principal investigators (or equivalent) at academic and nonprofit research institutions can sign up, then add the researchers in their group. Over the coming months, we plan to extend the program well beyond the initial 10,000 seats.
Learn more: claude.com/programs/team-pla…
Matteo Fischetti retweeted
Research Handbook on Transport Modeling e-elgar.com/shop/gbp/researc…
Matteo Fischetti retweeted
🗣️We have invited Giovanni Di Liberto (@diliberg), from @tcddublin to speak us about investigating auditory #cognition with natural speech and music.
We are eager to learn more about his work!
Discover more about it here👇
bcbl.eu/es/noticias/ponente-…
Matteo Fischetti retweeted
Subject to: Katya Scheinberg (Georgia Tech), youtu.be/7hWcofzu-Tc, #orms #optimization #WomenInOR @GeorgiaTech @GeorgiaTechISyE @math_opt @INFORMS @siam @TheSIAMNews @INFORMS_WORMS
Matteo Fischetti retweeted
I sat down with @JonHernandezIA in Madrid to discuss the growing risks and impacts of AI and the urgent need to improve our social, political, and technical safeguards.
Thanks for an excellent conversation!
📁 Yoshua Bengio, Turing Award winner and AI pioneer, says AI agents are learning to pursue complex goals without human supervision.
The danger is that once systems optimize for outcomes over instructions, they begin creating unethical subgoals on their own.
As agency scales into robotics and infrastructure, losing control stops being theoretical.
Matteo Fischetti retweeted
I write about this in more detail in a blog post with a guest contribution from Isaac Rajagopal, a student at MIT on whose work ChatGPT built, who gives his assessment of the level of mathematical ability displayed by the model.
gowers.wordpress.com/2026/05…
