Major update from Xaira Therapeutics!

Worrying results in brain scan studies as a result of ChatGPT!

First position paper I ever wrote. "Beyond Statistical Learning: Exact Learning Is Essential for General Intelligence" arxiv.org/abs/2506.23908 Background: I'd like LLMs to help me do math, but statistical learning seems inadequate to make this happen. What do you all think?

What if AI isn’t about building solo geniuses, but designing social systems? Michael Jordan advocates blending ML, economics, and uncertainty management to prioritize social welfare over mere prediction. A must-read rethink. arxiv.org/abs/2507.06268…

Distributional diffusion models with scoring rules at #icml25 Fewer, larger denoising steps using distributional losses! Wednesday 11am poster E-1910 arxiv.org/pdf/2502.02483 Alexandre Galashov Valentin De Bortoli Guntupalli Guangyao (Stannis) Zhou Kevin Patrick Murphy Arnaud Doucet

📢 New paper on Multi-Agent LLMs 📢 Our new paper presents Multi-agent-guided Leader Policy Optimisation (MLPO). We train a single leader LLM that steers a team of off-the-shelf agents to solve tasks. Detailed thread below 👇 🧵 arxiv.org/abs/2507.08960

Awesome to see this epic piece of work, led by Cecilia Clementi, finally appear in Nature Chemistry. The development of a general coarse-grained protein forcefield to describe folding, binding and conformation changes without solvent and all-atom, has been long anticipated!

Curious why such a flop compared to last year…

Excited to share that a scaled up version of Gemini DeepThink achieves gold-medal standard at the International Mathematical Olympiad. This result is official, and certified by the IMO organizers. Watch out this space, more to come soon! deepmind.google/discover/blog/…

Orthonormal updates appear to roughly double transformer training convergence with Dion paving tractability at the largest scale. Code: github.com/microsoft/dion Paper: arxiv.org/abs/2504.05295

Should we still teach machine learning classes? Let me justify my existence. argmin.net/p/just-when-i-…

Days in a work week: 5 Days in a month: 30 Total new submissions to arXiv in September: 26,646 arXiv editorial and user support staff: 7 someone who is good at science please help me with this. our team isn't sleeping. #openaccess #preprints

One of the coolest datasets to come out on Hugging Face recently is a collab between University of Oxford and Mila - Institut québécois d'IA. 🧑‍🔬 ManyPeptidesMD Huge molecular dynamics with 4.3 ms across 21,700 peptides. All info on the data gen available. Comes with a model, too. 👏 charliebtan

One of the most important papers of the year.

A key claim here arxiv.org/abs/2511.05963 is that next token prediction has no inherent preference for a heliocentric Copernicus theory en.wikipedia.org/wiki/Copernica… over a geocentric Ptolemy en.wikipedia.org/wiki/Ptolemy#A… theory of observations. Predicting the next latent fixes that.

An LLM-generated paper is in the top 17% of ICLR submissions in terms of average reviewer score, having received two 8's. The paper has tons of BS jargon and hallucinated references. Fortunately, one reviewer actually looked at the paper and gave it a zero. 1/3

“One of the continuing scandals in the physical sciences is that it remains in general impossible to predict the structure of even the simplest crystalline solids” (John Maddox) OXtal is a new generative model addressing this holy grail problem Blog: oxtal.github.io

Particularly eloquent essay!

Strawberries? From Dyson?! Yes, the same Dyson that reimagined the vacuum cleaner… is now reimagining farming. And it’s brilliant. When I think of Dyson, I picture futuristic fans — not vertical farms. But here we are. The innovation? Dyson has developed a high-tech

Joint Statement: Apple and Google have entered into a multi-year collaboration under which the next generation of Apple Foundation Models will be based on Google's Gemini models and cloud technology. These models will help power future Apple Intelligence features, including a