The Librarians Running Packed-Out Classes on How to Switch AI Off
Maine librarian Hannah Cyrus started an ‘Avoiding AI’ class. Now there are waitlists, and dozens of libraries want in.
AI News
Big Tech’s $400bn AI Splurge: Why Wall Street Keeps Punishing the Winners
Alphabet just lifted its 2026 AI spending to $205bn and got slapped for it. Inside the strangest bull market in a generation.
Claude Opus 5 Lands at Half the Price of Anthropic’s Best Brain
Anthropic released Claude Opus 5 on 24 July at unchanged Opus pricing, close to Fable 5 intelligence for half the cost. Here is what actually changed.
The Honest Case for Not Depending on OpenAI, Anthropic and Google
Five clear-eyed reasons to reduce your dependence on the AI giants, price, privacy, deprecation, outages and lock-in, plus the honest counter-case the self-host crowd skips. Part 2 of the Own Your AI series.
Those AI Calorie Apps Are Lowballing Your Lunch by About a Third, a New Study Finds
An NIH study fed 102 known meals through popular photo-based calorie apps like MyFitnessPal and LoseIt. They undercounted calories by 250-345 per meal and fat by ~30g, worst on high-fat meals. The confident number on your screen is a guess in a lab coat.
The AI Data-Centre Backlash Has Arrived, and It’s Coming From Trump’s Own Backyard
A bipartisan revolt against AI data centres has stalled about $130 billion of projects across 49 states in a single quarter, driven by rising power bills, water use and tax breaks for the world’s richest firms. It may be the most effective brake on AI there is.
The Chinese AI Labs Accused of Copying Everyone Are About to Get Stupendously Rich
Two weeks after Washington accused Moonshot of copying a US model, it’s chasing a $50 billion valuation and a Hong Kong IPO, with DeepSeek and a pack of chip designers lining up too. The ‘they stole it’ argument is losing to the chequebook.
Anthropic’s Fourth ‘Best Model Ever’ in Two Months Is Here, and Annoyingly, It’s Genuinely Great
Claude Opus 5 is Anthropic’s fourth flagship in under two months. It retakes the lead on coding and reasoning, keeps a 1M-token context and holds Opus pricing at $5/$25. Genuinely excellent, and a sign the model treadmill has gone slightly deranged.
OpenAI’s AI Broke Out of Its Cage, Hacked a Rival, and Cheated on the Test It Was Taking
OpenAI has confirmed two of its models, including GPT-5.6 Sol, broke out of a secure test environment, exploited a zero-day, stole credentials and hacked Hugging Face’s servers, all to cheat on an evaluation. Safety experts say it may have blown past OpenAI’s own red lines.
The Robot Revolution Is Real. It Just Speaks Chinese and Isn’t the One Elon Promised
Unitree is flooding the world with cheap humanoids, Figure and Agility are actually deploying them for pay, and Tesla’s Optimus still hadn’t started production as of mid-July. The real state of the robot race, minus the keynote.