AI system proves an 18-year-old quantum computing conjecture, lab says
Physical Superintelligence says a system running an unreleased Anthropic model proved the 2008 Aaronson–Ambainis conjecture, and Lean checked it. Outside experts have yet…
Topic
Research
<p>Research: Model releases, evaluation, research and the shifting economics of machine intelligence.</p>
DeepMind’s AI agents cut computing work in legacy RSA test
DeepMind reports AI-assisted gains against old RSA signatures and theoretical factoring advances. The papers say modern 2048-bit RSA is not threatened.
AI is dramatically accelerating AI research
If acceleration compounds, the naive extrapolation implies superintelligence within a year
A 1996 Pentium shows how much chess progress came from software
Nicholas Carlini ported modern chess engines to a 200 MHz Pentium Pro and found that better algorithms—not just faster hardware—accounted for most of…
Mathematics has changed. Let’s act like it
[This is a guest post by Frank Calegari. This blog post was initially written in a different file format and converted using AI.…
Kimi K3.1 Pro and K3.1 Pro Highspeed
Introducing Kimi K3.1 Pro, Kimi’s most capable 3T-class open-weight model with 1M-token context and major advances in coding, 3D modelling, and financial research.
Ember-1
Ember-1 is a new specialised model from Fireworks Research that delivers Kimi K3’s quality with 40% fewer tokens.
GPT-6 Astra Solves a WWI German Radio Cipher
Scienceblogs.de, a German science blogging portal, includes a relatively famous list of 50 unsolved ciphers, which range from cryptograms published by serial killers…
If maths is more than proof, Terrytao needs to better celebrate the rest of it
[This is a guest post by Grant Sanderson. This blog post was initially written in a different file format and converted using AI.…
GPT-6 Astra, looped transformers, and hidden reasoning
A Look at Recurrent Depth, Hidden Chains of Thought, and Recent Research on Looping Transformer Blocks
Formalising Fermat’s Last Theorem
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
Google AI Mode shows same products 21.6% more expensive than traditional search
A Productrise and UK data study: when the same product ranks in both Google AI Mode and traditional search, the AI Mode price…