IS WHAT I WOULD SAY IF NO ONE SHOWED UP. But 7,000 of you did. In twelve days. I built Clief Notes because I got tired of watching good people get sold bad AI advice. You all showed up because you were tired of it too. And now we've got something worth protecting here. Every single one of you who posted a question, dropped a comment, shared your work, or just lurked and learned: thank you. You're the reason this thing has weight to it. Here's what's coming: The Foundation course is live and growing. The Vault and Drawing Room are filling up with people who want to go deeper. High Tea kicks off in 10 days. And we're just getting started on the build-out. This community runs on one rule: we build things. We learn the fundamentals, we understand what goes where, and we make stuff that works. If you're here for that, you're in the right place. Bring your friends. Bring your weird project ideas. Bring the thing you've been stuck on for three weeks. We'll figure it out. More announcements soon. Stay loud. — Jake

If we haven't met or read any of my posts here. Me: Computer user before PC's were available. Built my 1st one in 78 with a soldering pencil, think rotary phones. Now WebDev and Marketing consultant. Why your here. I develop full stack websites with PayloadCMS after switching from years of WordPress. What I'm sharing here is a PayloadCMS Project Development workspace template. To use it's as simple as cloning/forking a repo and including the folder in your local dev project. Once opened in your ide of choice, run setup. Key Benefits of This Approach - Consistency: Repeatable process ensures quality across projects - Clarity: Explicit stage contracts prevent misunderstandings - Efficiency: Reusable components and patterns reduce development time - Quality: Built-in validation checkpoints catch issues early - Knowledge Retention: Process captures organizational learning - Scalability: Approach works for both simple sites and complex applications 1st it asks some questions: I'll help you set up your PayloadCMS project development workspace. Please answer all of the following questions in a single message: 1. Project name: What is the name of your PayloadCMS project? (Default: my-payloadcms-site) 2. Website type: What type of website are you building? (Options: Blog, Business site, Portfolio, E-commerce, Documentation, Other) 3. MVP stage: Do you want to include MVP development as an optional precursor stage? (Default: yes) 4. Deployment target: Which deployment target do you plan to use? (Options: DigitalOcean, Vercel, Cloudflare, AWS, Other) 5. Database: What database will you use with PayloadCMS? (Options: PostgreSQL, MongoDB, SQLite, MySQL) 6. Vercel best practices: Do you want to bundle the vercel-react-best-practices skill for frontend optimization? (Default: yes) 7. SEO configuration: Do you want to include SEO configuration in your PayloadCMS setup? (Default: yes) 8. Search functionality: Do you want to include search functionality in your PayloadCMS setup? (Default: yes)

I would love to get some feedback on my situation/dilemmas. I had accumulated some know-how in defining and encoding the workflow, context, structure, constraints, etc., for different tasks without using tailored personas/instructions for projects within the main AI chat apps. Then I thought that most employees don't have the time or energy to dive deep and dig through layers of AI influencers, prompt packs, etc., to actually optimize their work beyond one-time conversations or simple tasks, and so I could bring some value. Went all in on finding, learning, and designing the best ways to achieve that. As Opus 4.6 and Codex 5.3 came out, I dove into the "agentic" area and its possibilities, where I feel my experience has whole new potential, as there is only so much that can be done within projects and instructions, and they allow for hardcoding some deterministic parts. That led me to realize that, in the current state, I can leverage my somewhat shallow understanding and insight into different areas, as models can compensate for my lack of technical background. So I was able to create a website with an integrated reservation system for my wife's lessons. And, at least in my opinion, it escapes the generic feel of most AI websites. So I defined some workflows and principles to achieve similar results faster for other potential clients. What I am trying to figure out are 3 things: - Whether to keep exploring more complex solutions or to focus on improving people's workflows in the everyday layer of the main apps through instructions/projects. Since I feel most people are and will continue to use them as their primary layer. - How to deliver the value in the best way, as I am not much of a salesperson and have an ick about selling courses or teaching for the sake of it. - I feel the value is there, but it is hard for me to define it clearly and position myself. And it all kind of falls under generic categories of AI workflows, consulting, specialist which are oversaturated

You just saw how context changes per task. Vote below, then tell us in the comments: what two areas of your work did you set up, and did Claude behave differently in each one?

For those of you who are in academia and would like to cite any of my research. I just pushed my ICM paper and had it submitted to preprint. You'll notice that the working title is called ICM but I refer to the workspaces MWP, I am pushing an update right now removing all mention of MWP but Model workspace protocol was the original name I had called it. However I think Interpretable Context Methodology makes more sense for what is actually happening. Enjoy! https://arxiv.org/abs/2603.16021