Practical ChatGPT tip I wish I'd thought of earlier: I've started replacing a lot of the individual SourceDocs in my projects with organized ZIP archives of the underlying files. I can keep substantially more project context available to ChatGPT while making the Sources section much cleaner. Instead of maintaining 20 different context documents, I can organize the actual material into a few logical archives and keep adding to them. I originally created SourceDocs partly because I needed to compress a project into something AI could work with. I'm realizing I can increasingly just preserve the real source material and organize it well. If you use ChatGPT Projects for serious long-running work, this is worth experimenting with.

Most people think “building in public” means marketing. Post updates. Show progress. Grow an audience. Get attention. That can happen. But I think the more important benefit is learning. ╭────────────────────────────────╮ │ 🧭 THE IDEA ╰────────────────────────────────╯ When you build something privately, it is easy to move fast and leave your thinking scattered. You make a decision. You fix a problem. You discover something useful. Then you move on. When you build in public, you have to stop and explain what happened. That changes the work. You start asking: What did we actually learn? Why did we make this decision? What failed? What would another builder need to know? What artifact could preserve this? ╭────────────────────────────────╮ │ 📜 THE GUILD PRINCIPLE ╰────────────────────────────────╯ BUILD → EXPLAIN → REFLECT → IMPROVE → PRESERVE The act of explaining your work forces you to understand it more clearly. A screenshot can become evidence. A mistake can become a lesson. A design decision can become a SourceDoc. A working prototype can become a demonstration. A confusing problem can become tomorrow’s quest. ★ THE PROJECT DOES NOT ONLY PRODUCE A PRODUCT. ★ THE PROJECT CAN PRODUCE CURRICULUM. ╭────────────────────────────────╮ │ 🌍 TODAY’S EXAMPLE ╰────────────────────────────────╯ We are beginning to build WorldOS in public. The project itself is ambitious: make publicly available world information easier for machines to retrieve and use while preserving what each source communicated and where it came from. But WorldOS is also becoming something else. A learning trail. As we make architecture decisions, choose sources, build connectors, hit problems, publish code, and change direction, those moments can become artifacts for the Guild. Someone following the project does not need to build WorldOS themselves. They can learn from the process of building it. ╭────────────────────────────────╮ │ ⚔️ TODAY’S QUEST ╰────────────────────────────────╯

You’ve already been training the dataset One thing I think we’ve been missing when we talk about “training your AI”: You’ve already spent years creating data about yourself. Spotify has your listening history. YouTube has your watch history. Google Photos has years of places and memories. Your calendar knows how you actually spend your time. So a good first AI project might simply be: Get some of your own data back and see what you can build with it. For example: Spotify history → AI analyzes your actual music taste → n8n checks upcoming concerts → you get a message when someone you’d probably love is playing nearby. That one little project teaches data, connectors, automation, AI reasoning and personalization without starting with a lecture on any of them. 🧪 Guild experiment: Which account would you mine first? And what would you want your AI to learn from it?

Most people treat AI conversations as disposable. We’re beginning to treat them as a learning history. Today’s post is actually an example. I asked AI to look across our previous conversations and find experiences that had already taught us something. It surfaced moments like: - I forgot how my own website worked, found the old conversation where we built it, and reconstructed the project from that saved state. - A valuable conversation disappeared. The main idea survived—but the exact wording, objections, examples, and turning points did not. - Across several projects, the same pattern kept appearing: real activity becomes source material, which becomes a lesson, workflow, or new system. That changed my mental model. A single prompt tells AI what you’re thinking now. Continuity lets AI compare now with before. The difference between those two states is where many of the best lessons are hiding. ⚔️ Today’s quest Find one older AI conversation about something you were learning, building, or deciding. Ask: “Read this as a learning historian. What did I believe at the beginning? What changed? What evidence caused that change? What could this teach someone else?” Share one result: Old topic: Recovered lesson: What changed in my thinking: Remove anything private before posting.

Can you explain how 1s and 0s become TikTok? That might sound like a joke, but I think it’s one of the most important questions we can ask right now. Most of us use computers every day without really understanding what is happening underneath the screen. Somewhere between: 1s and 0s → electricity → logic gates → processors → operating systems → networks → cloud servers → databases → software → TikTok …there is an enormous gap in our mental model. And now AI is giving us the ability to build increasingly powerful things without necessarily understanding that stack. That is incredible. It is also exactly why I think computer science fundamentals are becoming more important, not less. You do not need to become a computer engineer. You do not need to memorize binary or write assembly code. But if AI is going to become one of the most powerful tools in your life, I think you should understand the machine you are directing. So here’s today’s Guild question: Could you explain, even roughly, how a 1 and a 0 eventually become an app like TikTok? If the answer is no, good. That means we found a quest. ⚔️ That’s exactly the kind of gap we’re going to start closing together. #AIInnovatorsGuild #AILiteracy #ArtificialIntelligence #ComputerScience #CloudComputing #AIAgents #LearnAI #FutureOfEducation #TechEducation #DigitalLiteracy #AIEngineering #LearnByBuilding