Welcome to the AI Untapped Workshop. This is the room behind the newsletter and the YouTube channel. The whole point: cut through the AI noise, faster, together. Here's what lives here: → One new tool + prompt every Sunday. Posted in the feed. Copy, paste, run it. → The Field Notes: five short PDFs covering local AI setup, prompt patterns, the skip list, the maker stack, and the Sunday routine. Find them in the Classroom tab (top nav) or scroll the feed today, where each one is posted separately so you can ask questions in the comments. → An open thread where you drop the AI tools you're testing and what's actually working. → Behind-the-scenes context I don't put on YouTube. What I'm asking from you: → Post what you're testing. Wins and failures both. The room learns faster when we're honest. → No pitching. No "I built an agency, DM me." This is a workshop, not a marketplace. → If a prompt works, drop the result in the comments. If it breaks, drop the error. I'm here daily. Drop a hello below and tell me one thing you're trying to figure out with AI right now. I read every reply. — Dusty

Several universities have stopped trusting the software that claims to spot AI written text. Yale and Johns Hopkins have restricted these tools. The University of Waterloo turned Turnitin's AI detection off across every faculty. Yale's teaching centre said the false positive rate is too high to use as proof in an academic honesty case. A false positive is when the tool flags writing as AI even though a person wrote it. That has always been the weak spot. These tools do not read your work and recognise a machine. They measure how predictable the writing looks, and predictable reads as machine. So the student who writes in clean, simple sentences gets flagged, and the one who rambles slips through. This reaches past universities. Some employers run the same checks on job applications, cover letters and written tests. If the institutions that made detection software normal are now pulling it out, that is worth knowing before someone runs your writing through one. Have you ever had something you wrote flagged as AI? What happened next?

Quick heads-up for everyone: **Claude Fable 5 is available again as of July 1, 2026.** If you've been waiting on it, you can start using it today. (until the end of July 7th) Here's the short version of what happened and what it means for you. -> What happened Fable 5 (and its sibling, Mythos 5) first launched on June 9. A few days later, on June 12, the U.S. government placed **export controls** on both models, which forced Anthropic to suspend access for everyone. The reason: a report showed a way to get Fable 5 to help identify software vulnerabilities. Anthropic's own testing later found that plenty of weaker, already-public models (including older Claude versions and other companies' models) could do the exact same thing — so it wasn't a unique risk. Still, they tightened things up. **As of June 30 the export controls were lifted**, and Fable 5 came back the next day. -> What's different now Anthropic added a stronger safety layer before bringing it back: - A new **safety classifier** that blocks the specific technique from the report in **over 99% of cases**. - If one of your requests to Fable 5 ever gets blocked, it gets **automatically rerouted to Opus 4.8** so you're not left stuck. - One tradeoff to know about: the new safeguards are a little "trigger-happy." You may occasionally see a normal coding or debugging request get flagged by mistake. Anthropic says they're actively tuning this down. -> What this means for your access - **Where you can use it:** Claude.ai, Claude Code, Claude Cowork, and the Claude Platform (API). - **Usage limits (Pro, Max, Team, and select Enterprise plans):** Through **July 7**, Fable 5 is included for up to **50% of your weekly usage limit**. After July 7, continued access runs on **usage credits**. - **Cloud platforms:** AWS, Google Cloud, and Microsoft Foundry access is being switched back on "as quickly as possible" — not instant, so check your provider if that's how you access it. -> Bottom line

Most people treat Claude like a search engine. Ask a question, copy the answer. That leaves most of the value on the table... Try this instead. Whenever Claude hands you something, reply with: "What are the three weakest parts of this and how would you fix them?" It forces the model to critique its own work, and the second version comes back sharper almost every time. I run it on emails, scripts, business plans, code, anything I care about getting right. One extra trick: ask it to rewrite only the weakest part, so you keep whatever already worked. Try it on something today and tell me what it caught.

Quick heads up. Fable 5 came out on June 9, got pulled two days later when the US put export controls on it, and as of this week it's live again for everyone. If you're on Pro, Max, or Team, you get it for up to half your weekly usage through July 7, then it runs on usage credits after that. So this is basically a free week to test the most capable model Anthropic has shipped. Here's why it matters for us. This is the model you point at a messy job and walk away from. Drop it into Claude Code or Cowork and it will plan the work, break it into stages, hand pieces off to sub-agents, and check its own output over hours or even days. You're not sitting there babysitting a chatbot. If you've been putting off a project because it felt too big to start, this is the week to throw it at Fable 5 while the usage is cheap. What's the first thing you're going to test it on? Drop it in the comments.