1. Five whys: stop your AI repeating the same mistake

    A post-mortem prompt for the end of a bad AI chat: it asks why five times and gives you one rule to check before you paste it into your AI's instructions.

    5 min read Interactive

  2. Shared project instructions: one rulebook for your team's AI

    Shared project instructions keep your team's ways of working and house rules in one document every AI reads: a shared Claude Project, ChatGPT project or Gem.

    4 min read Interactive

  3. What to do when Claude says it's fixed, but it isn't

    A new study shows why an AI's fix can match the examples it was checked on and still fail on new ones, and how to test a fix yourself before you rely on it.

    3 min read Interactive

  4. Claude Enterprise inference hooks: what they catch and miss

    Claude Enterprise inference hooks send each prompt, and optionally each tool call, to your security server for an allow or deny. What they check and miss.

    5 min read Interactive

  5. GPT-6 Cyber found 81 Firefox flaws: what to ask suppliers

    OpenAI says its new model found 81 high-severity Firefox flaws. Ask suppliers if they bundle software like PHP, and what protects your data until it's fixed.

    6 min read

  6. Use your Claude, Codex and Cursor usage limits before they reset

    Each plan's allowance resets on its own schedule. Keep a list of donkey work for when you have plenty left, and save a limit reset for when you run short.

    3 min read Interactive

  7. Create an AI skill from frameworks like the Eisenhower matrix

    Start from a framework that hasn't changed in decades, like the Eisenhower matrix. Give the AI its Wikipedia link, then answer its questions one at a time.

    6 min read Interactive

  8. Keep your site out of Google's AI Overviews and AI training

    A Search Console setting keeps your site out of Google's AI Overviews; robots.txt rules keep it out of AI training. Your site is in both until you change them.

    6 min read Interactive

  9. Skills, then agents, then context

    Google is turning Gems into skills. If a big AI job overwhelms your team, break it into simple steps, write each as a skill, and put an agent in charge of running them.

    4 min read Interactive

  10. Loyal to your work, not your model

    Google is turning Gems into skills. The AI tools you rely on will keep changing, so keep your prompts, skills, logs and chat summaries outside the chat, where you own them.

    4 min read Interactive

  11. Whether we call it a hack or not

    An AI model found a published way into Georgia's vote-count data and used it. Whether we call it a hack or not, a way in that hasn't been properly plugged can lead to more.

    4 min read Interactive

  12. Steam saved a 60 fps cap for 112 games. The Performance panel still showed 144

    An AI agent changed a setting on my Legion Go, and Steam's config file showed it saved. A screenshot of the settings panel showed it hadn't taken effect. Anything new or recently adjusted needs oversight.

    2 min read

  13. Give a shared AI assistant a weekly checkup, and keep what it adds terse

    Once a week, ask the assistant your team shares what would improve its instructions. Keep any rules it adds succinct, and give the bot two human owners.

    2 min read

  14. Ask your AI to save its work as files

    I've kept a daily log in a file since the early days of using models, so I can switch models or start a new chat and pick up where I was. Ask for your work as files, and keep them in named folders.

    3 min read

  15. Plan big, do small

    A big model should plan the work and find the tools, libraries and free databases that already exist. Smaller models, or programs on your own computer, do what I like to call the donkey work.

    5 min read

  16. Logo.dev Returns a Company's Logo From Its Web Address, and Its MCP Server Lets Claude Fetch It

    Logo.dev returns a company's logo, colours and brand details from its web address. Here is what it does, how to connect it to Claude, and what to use it for.

    3 min read

  17. When Gemini Runs Claude, What Are You Buying?

    Google’s workplace agent can use models from more than one provider. I want to know what happens to my instructions, files and work history when I switch.

    3 min read

  18. An Agent Needs a Stopping Point

    A failed connection or a missing file should have a defined outcome. My inbox cleanup and Anthropic’s latest report offer a practical way to specify it.

    2 min read

  19. The 3,848-Message Sieve

    What It Takes to Let an Agent Clean Your Inbox

    3 min read

  20. OpenAI's DevDay is a pile of announcements, for now

    The Decisions API looks good, a few things need testing, and Pro 500 set the cat amongst the pigeons.

    1 min read