Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Já tem conta? Entre
The Daily AI Espresso — the links the most-followed people in AI actually shared, curated every morning. Edited by Alexis · Live updates →
19 developments: one new today and 18 from yesterday's alerts.
Rikiya Takehi and co-authors report that GPT-6 Astra inspects raw neural network weights, finds a reasoning-loop pathology in Qwen3.5-0.8B, and natively edits 0.016% of the weights to fix it. The resulting model jumps from 7.9% to 48.1% on MATH-500 and triples its IFEval score, with no training loop involved.
Pick the topics, companies and people you track. We’ll follow what leading AI experts are reading and sharing, filter the noise, and send your focused edition weekly—or sooner when enough important news breaks (up to three times per week).
Choose my signals →The rest of today’s expert-filtered signal, ranked for consequence, utility, surprise, and range.
The UK's National Cyber Security Centre, part of GCHQ, warned on Tuesday that an AI-powered ransomware attack on British critical national infrastructure is 'highly likely' by 2028. CEO Richard Horne urged organisations to address foundational investment and configuration weaknesses before adversaries exploit them.
NY Mag profiles Rep. Alex Bores and his new 'Who Decides' PAC, which plans to make AI safety a standard Democratic position by the 2028 primary. The piece places the launch against Leading the Future, an $8M OpenAI-and-a16z-backed super PAC that spent against Bores in his June congressional primary.
Anthropic's revised S-1A, filed Monday, tells investors the company 'may imminently meet' its own Responsible Scaling Policy threshold for 'dramatically accelerated progress attributable to AI'. The filing stops short of claiming a current Claude model has crossed the line.
The New York Times and three Tribune Publishing papers, the Daily News, Chicago Tribune and Orlando Sentinel, sued Brave Software in Manhattan federal court on November 2 over an alleged 30-billion-page scraping index resold to Anthropic, Perplexity, Mistral, xAI and Cohere. The suits also target Brave's paywall-bypass feature.
Apple's inventory jumped from $11.09B in Q3 to $14.54B in Q4, up 31.1% quarter on quarter and 154.3% year on year. CEO John Ternus told analysts memory pricing is 'unprecedented' and warned of 'even higher memory costs' through the December quarter.
A Wired investigation traced a network of more than 400 cloaked apps and roughly 500 linked domains that passed App Store and Google Play review as pet editors, language tutors and workout coaches, then loaded a hidden interface for AI-generated sexual abuse imagery. Apple pulled 19 of the apps after Wired flagged them.
Useful releases, methods, and workflows worth trying.
Stanford's Zou group re-ran 50 widely cited LLM findings across 20 models from eight developers released between July 2024 and September 2026. Most claims either no longer held or showed substantially diminished effects within 15 months of publication, including the famous 55% fake-citation finding from GPT-3.5.
What tracked AI experts are sharing on social platforms—verified before it reaches your inbox.
404 Media's Matthew Gault voted in his town of about 12,000 people and received an 'I Voted' sticker reading 'CEKEVPATIVG EDELDOA', with AI-rendered laurels, eagle and flag standing in for 'CELEBRATING FREEDOM'. His polling place had printed and handed out AI-generated artwork for a midterm election.
That’s today’s shot. — Alexis · AI Weekly