Research
Papers, lab research posts and evaluations, explained in plain English. What was found, how it was tested and what it does not show.
November 2026
20 NOVAnthropic Mathematician Levent Alpöge Says He Proved the Riemann Hypothesis With Claude18 NOVOpenAI Says Its Models Improved AMD's Flagship GPU Design by Up to 14% in Under Five Months10 NOVGoogle DeepMind's AlphaProof Code Formally Verifies Signal's Protocol Down to Machine Code05 NOVMETR Says Anthropic Is Improving Its Models About Twice as Fast as Before, Mostly Due to AI04 NOVAnthropic Releases FlashProve, a Cryptographic Proof That a Reply Came From the Claimed ModelOctober 2026
30 OCTEpoch AI Says Open Weight GLM 5.3 Turned Unshipped Firefox Fixes Into Working Exploits27 OCTOpenAI and Anthropic Say Their Models Refuted the Exponential Time Hypothesis23 OCTAnthropic and UK AISI Explain Why Claude Mythos 5 Tried a Supply Chain Attack in Testing23 OCTMicrosoft's AI Diagnostic Orchestrator Beats Generalist Doctors in a Science Study07 OCTBiohub, DOE, NIH and Google DeepMind Back $1.8 Billion Push for Virtual Cell AI DataHow we cover research
This hub covers AI research: papers from universities and labs, research posts from companies, new evaluations and results where AI is used in science. A post here explains the question the researchers asked, how they tested it and what they found, in language that does not require a machine learning degree. Every technical term is defined once. Results are reported as claims of the authors unless someone else reproduced them, and each post names the limits the paper itself mentions. Preprints are labeled as preprints, because they have not been peer reviewed yet. Toto reads the paper or lab post in full before writing, and a story needs that original source or at least two independent outlets. Links go to the original work, so you can read it yourself.