Super exciting paper by the groups of Jan-Åke Larsson & Marek Żukowski (two top-experts in Bell's theorem): arxiv.org/abs/2508.19207 They analysed our recent paper on a very strange nonlocal interference effect (science.org/doi/10.1126/sc…). They found that while the inequality we

Also our experiment was covered by a video today by Sabine Hossenfelder: youtube.com/watch?v=QBO531…

Our paper: science.org/doi/10.1126/sc…

This sounds amazing: How do Humans and LLMs Process Confusing Code?arxiv.org/abs/2508.18547 by Abdelsalam Norman Peitek Maurer Mariya Toneva, Apel Max Planck Institute for Software Systems Saarland Informatics Campus We (humans & machines) are both surprised by confusing code 🙃🤖👽

We have press! A fun and inspiring piece to be thinking about :D newscientist.com/article/mg2673…

I am joining Axiom Math, a seed-stage startup on AI for maths. I will lead discovery: AI for advancing math research. 6 years after Deep Learning for Symbolic Maths, our first paper with Guillaume Lample @ NeurIPS 2024, I am proud of the field's progress, and excited about what comes next.

Can artificial intelligence predict future research directions in quantum science? Felix Frohnert and I discuss this question with Hamish Johnston Physics World: physicsworld.com/a/artificial-i… About our recent Machine Learning: Science and Technology paper: iopscience.iop.org/article/10.108…

Using AI to advance our understanding of fundamental physics is the dream. Excited to see our latest AI model 'Deep Loop Shaping' help LIGO and Caltech detect the gravitational waves of intermediate-mass black holes better! Published in Science Magazine

Tutorial on how to contribute to PhysLean through golfing proofs🏌 : 📽️ youtube.com/watch?v=QL-mjk… This is a good way to get involved if you are new to Lean but want to get stuck in. #physics #LeanLang

Obituary: Rainer Weiss (1932-2025) Nobel laureate who pioneered the technique that detected gravitational waves go.nature.com/4m46Gu4

Today LHCb presented a new measurement of the decay of a B meson into a K* and a μ+ μ- pair Previous results disagreed with the Standard Model prediction for the branching fraction and angular distribution The new measurement has almost twice as much data and still disagrees!

Today we're announcing Gauss, our first autoformalization agent that just completed Terry Tao & Alex Kontorovich's Strong Prime Number Theorem project in 3 weeks—an effort that took human experts 18+ months of partial progress.

Can you help PhysLean make better documentation for the classical harmonic oscillator? If so take a look at this link where you can review what we currently have and make suggestions for improvements! 🔗 github.com/HEPLean/PhysLe…

The clearest gravitational wave signal ever seen provides strong evidence for Hawking’s area law This law is best understood via Bekenstein-Hawking: a black hole’s entropy is proportional to its surface area. Entropy never decreases, so a black hole’s area can never decrease

Blog post: "Actually, you can't test if quantum uses complex numbers" algassert.com/post/2501 I doom the concept of that 2021 Nature paper by showing how to compile any distributed quantum protocol into real-only gates while preserving locality.

This is a surprising twist of events: the much celebrated quantum info Nature-paper from 2021 on falsifyinig real-numbered quantum physics seems to rely on the (hidden) assumption that no entanglement is used (and if entanglement is included, as it is in real world, the result

An honour to feature on the homepage of DIE ZEIT today, for the first in their new series "Just One Question", launching across print, online, audio, video, & newsletter. 🙏🏽 Jochen Wegner for the conversation, & Marie Luise Schreiter for help with the translations zeit.de/wissen/2025-09…

Would you let AI cheat for you? Our new paper in @nature.com, 5 years in the making, is out today. nature.com/articles/s4158…

WOW these are some cool projects: Formalizing Modern Theorems. "...dataset of hundreds of formalized statements of theorems from top journals, such as the Annals of Mathematics..." LeanTutor. "...reliably auto-formalize undergraduate students' proofs."

🚨New Meta Superintelligence Labs Paper🚨 What do we do when we don’t have reference answers for RL? What if annotations are too expensive or unknown? Compute as Teacher (CaT🐈) turns inference compute into a post-training supervision signal. CaT improves up to 30% even on