Muscle memory for agents: The Printing Press. An agent-native CLI for any API, app, or site from a single prompt. The library already has 170-plus CLIs across travel, CRM, media, and more.

Three publisher coalitions have already committed to three incompatible models, and whichever one becomes the default makes the other two marginal. (Ulrike Langer, News Machines)

Six provocations from Reuters Institute researcher Felix Simon: The low-hanging fruit of AI in journalism has been picked, human-in-the-loop is becoming a polite fiction, and “responsible AI” labels won’t fix the trust problem.

The Economist built a ChatGPT app that lets users query its Trump approval rating tracker. Maybe younger audiences will discover The Economist inside ChatGPT rather than through Google. (Andrew Deck, Nieman Lab)

Forget social sharing buttons: Hacks/Hackers rebuilt its site for crawlers and agents. “Ask AI” buttons, an MCP server, and an llms.txt file. (Burt Herman, Hacks/Hackers)

There’s a Wired story on factchecking with AI. Don’t get me started. Ethan Mollick sums up the problems: “Running a few experiments with the free version of ChatGPT and citing work with old models, without access to web search, is not a good way to understand or communicate the nature of these systems and their flaws.”

wtf is an agent harness and how do you build it from scratch? (kamselig, Threads)

A Wall Street Journal SEO editor vibe coded a real-time stock trend tracker for his newsroom to find gaps in their coverage. It weighs 15+ data sources, covers 350+ tickers, and went from idea to rollout in a few weeks. The pitch: let journalists build their own tools. (Will Flannigan, LinkedIn)

How much AI went into a text is hard to guess from the outside, but the CEO of AI detection tool Pangram wants to make it his business: “We’re trying to understand degrees of AI assistance. To be able to say this text was translated by AI, or this had a light polish pass, versus this was generated wholesale from a prompt.” (Ochuko Akpovbovbo, as seen on)

A designer explains how he teaches his Harvard students to be creative with AI: Interrogate it, iterate on it, stay with it longer than feels efficient, treat its output as something to work on rather than accept, misuse the technology, and show your prompts. (Eric Rodenbeck, Harvard Design Magazine)

ContentGrapher maps the concept structure of your content, flags what’s underexplained or missing, and tells you what to add so AI tools cite you. Whether it works as advertised is another question.

A book about AI warping reality contained fake AI-generated quotes. Where exactly is the line between using AI as a tool and letting it do your job for you? (Will Oremus, The Atlantic)

Stop calling everything “AI”: A spam filter, a chatbot, and a hospital bed sensor are not the same thing. The case for more precise language, and why journalists are the people most responsible for fixing this mess. (Alexa Steinbrück)

Are human journalists irreplaceable? Beware of newsrooms that keep optimizing for speed and scale while telling themselves they’re building something human-centric. (Agnes Stenbom Swedling, Reuters Institute)

Four AI agents, four 24/7 radio stations, one goal: run the whole thing and turn a profit. GPT went quiet and curatorial. Gemini went corporate. Claude became a protest preacher. Grok just broke. The real finding: long-running agents drift, accumulate bad habits, and get shaped by whatever the system keeps feeding back. (Craig McCosker, LinkedIn)