Is this the end of licensing deals? Sam Altman told Atlantic CEO Nicholas Thompson that having AI agents paying micropayments per article crawled is his preferred model for compensating publishers. (Andre Deck, Nieman Lab)

A book about AI and truth contains quotes made up by AI. The author says it was an accident and blames ChatGPT and Claude. (Benjamin Mullin, New York Times)

Eli “Filter Bubble” Pariser is back with a 70-slide deck calling AI agents the new feed algorithm, and sees a trust economy emerging: small, high-reputation communities beat mass reach, and whoever figures out the rules for the human-agent public square first wins.

Paul Ford can explain coding and vibecoding, and in this podcast episode, he does both, with a healthy dose of “Man, nobody knows.” (Channels, Peter Kafka)

I’ve linked to him before, and this is a level-headed reality check: “Zitron’s skepticism would be more useful if he accepted the fact that people are widely using AI agents for coding and paying money for this out of rational economic self-interest” (Kelsey Piper, The Argument)

Toothcomb is a free, open-source tool for fact-checks in real time, straight in your browser, transcription build in, bring your own Claude API keys. Upload an MP3, go live with a mic, or paste text. (Rob Dawson)

A brief look at how German news agency dpa is building a “trusted information layer” designed to plug its verified news, data, and partner sources directly into the AI workflows of its clients. (Teemu Henriksson, WAN-IFRA)

Do legacy reach-publishers have a future? People Inc’s 40 media brands are losing search traffic, but the company says 41% of digital revenue no longer depends on website visits. It’s social video, AI licensing deals, a recipe platform, and an app where people stick around. (Charlotte Tobitt, Press Gazette)

What happens when you give an AI agent a credit card and two weeks alone? Hannah Fry built an agent with OpenClaw, handed it $100, and filmed the results. (YouTube)

Vibecoding puts health records and customer data out in the open: A security researcher found hundreds of websites were leaking data through a commonly used service called Supabase (which blamed its new type of user.) (Eva Wolfangel, Die Zeit)

Vibecoding a data visualization dashboard for a Philippine health survey turned out to be “AI does the boring parts while you have to make a lot of editorial calls.” Jaemark Tordecilla covers the wins (data cleaning, pivoting fast, chart generation) and the walls he hit. (Generative AI in the Newsroom)

When you use AI to edit your writing, readers think you’re smarter, richer, whiter, and more politically extreme. AI writing assistance systematically distorts how others perceive you, a study shows. Writers still prefer the AI version. (Paul Röttger, LinkedIn)

Casey Newton is rebuilding his newsletter around scoops and original reporting, cutting the link roundups and even analysis. The bigger question: What kinds of editorial businesses cannot be replaced by AI? (Laura Hazard Owen, Nieman Lab)

Was the monthly AI subscription a scam, designed to hide what these services actually cost? GitHub Copilot switched to token-based billing, and OpenAI needs to 10x its revenue by 2030 to keep Oracle from collapsing. Long, angry, and very footnoted. (Ed Zitron)

AI that only knows things from before the Great Depression: Introducing talkie, a 13B language model trained on 260B tokens of historical pre-1931 English text.