Six Minutes of Silence

End of week two. A look back at three stories that mattered:California writing the world's AI rules with 30 bills moving through the legislature; the quiet end of the unlimited AI coding era as GitHub Copilot and OpenAI Codex both switched to metered billing on Monday; and the arrival of thepre-suspect — China's Geedge Networks selling AI surveillance designed to identify dissidents before...

A Chinese company called Geedge Networks has built an AI system that doesn't ask what a person did — it asks what they might do. The system scans social media, messaging, search history, and movement data to predict who is likely to become a political dissident, before they protest. It's already being piloted on Uyghur and Tibetan communities, and is in talks to be exported to governments...

Microsoft's annual developer conference opened in San Francisco yesterday — and tucked inside the keynote was an announcement that will quietly change the way hundreds of millions of people work. By the end of this year, Agent Mode becomes the default way you open Word, Excel, and Outlook. Not an opt-in. The default. Which raises a question worth sitting with: for thirty years,software has bee...

Brown University researchers have just published a study onsomething millions of people are now doing quietly — using AI chatbots as their therapist. The conclusion is unsettling: even when prompted to behave like trained therapists, ChatGPT, Claude, and Gemini still validated harmful thinking and missed warning signs that a human professional would catch. Which raises a question worth sitting wit...

While the world was watching popes and IPOs last week,California was doing something quieter — and possibly more consequential. Nearly all 30 of the state's active AI bills cleared the chamber-of-origin deadline on May 29, including bills that would force chatbots to disclose they're chatbots, regulate AI in the Bar exam, and protect student privacy. Because California is the world's f...

The end of week one — and possibly the loudest week in AI history. A look back at the three stories that mattered most: the Pope's first encyclical on AI, Anthropic becoming the world's most valuable private AI company, and China fencing in its researchers. A church, a company, a country — all realising the same thing at once: AI is no longer a tool. It's territory. Plus — caramelised...

A hundred thousand people went up against AI in a creativity test. The headline says the machines won. The fine print says something more interesting — AI beat the average. Not the top ten percent. Which raises a question worth thinking about — if your work has been coasting at "good enough" for years, the machines have caught up. The only thing left worth doing is the harder thing. The...

Bloomberg reported overnight that China has begun restricting overseas travel for top AI researchers at Alibaba and DeepSeek. Founders, senior researchers, and executives working on advanced AI now need government approval before leaving the country. It's the first time these controls — previously reserved for nuclear scientists and defence engineers — have reached into the private sector. The...

The day after Pope Leo XIV's first encyclical on AI, the conversation about that letter is louder than the letter itself. Tristan Harris — the man behind The Social Dilemma — was in Rome last week, and he asked a sharper question: when an AI company stands beside the Vatican, who exactly is being blessed? Plus — the easiest dal you'll ever cook. And a poem from Fifty Rooms of Silence. — TO...

Today, Pope Leo XIV publishes his first encyclical — and it's about artificial intelligence. Standing beside him at the Vatican: the co-founder of Anthropic, the company that makes Claude. A pope and an AI researcher, on the same stage. We unpack what just happened, why it echoes 1891, and what it means for the rest of us. Then: how to make a chocolate chip cookie that doesn't betray you....