A library of 42 copy-paste agent loop prompts for Cursor, Codex, and Claude Code. An agent loop is a task with a built-in check: do work, verify result, continue or stop. Some prompts are useful (“fresh-clone loop,” “devil’s advocate loop”), some are wishful thinking (“100% test coverage loop”), most land in between.
Google’s answer to AI agents fumbling around for context: put markdown files in a folder. The Open Knowledge Format is a simple spec for giving AI systems internal knowledge. (Sam McVeety, Google Cloud Blog)
The agentic divide is here: Well-resourced firms plug AI agents into their entire stack and scale infinitely. Everyone else gets high-friction, low-trust tools that can barely fill out a form. India’s government-led citizen agent programs are trying to close the gap. (Rina Chandran, Rest of World)
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)
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)
wtf is an agent harness and how do you build it from scratch? (kamselig, Threads)
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)
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)
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)
What if your AI agent doesn’t actually work for you? A field report from a Harvard summit on the agentic future of news: when personal AI agents become the primary gatekeepers, whoever shapes the agent’s “model of your intentions” becomes the most powerful editorial force in history. It’s scarier than the attention economy. (Lars Adrian Giske)
A young reader got so frustrated with the FT’s news product that she built her own using Claude Code. She’s paying for a subscription, she wants personalized news, and she couldn’t figure out how to get it. (Jodie Hopperton, INMA)
The vibe is Palantir cosplay: World Monitor by Elie Habib is a real-time “situational awareness” dashboard with threat feeds, geopolitical maps, and a three-stage AI classification pipeline. He calls it a “weekend hack.”
“Not ‘how do I get paid for my articles by AI’ but ‘how do I architect my knowledge so it can reach customers I’ve never had access to before'”: A three-layer AI monetization framework, with O’Reilly as proof of concept. (Florent Daudens, AI in the News)
Which AI to use: Free users don’t get the good AI, and even if you pay, you have to pick thinking mode. Once you’ve done that, the model matters less than the harness. For newsrooms: curated archive retrieval, a system prompt encoding editorial voice, CMS integration, verification. (Ethan Mollick, One Useful Thing)
