Repo Template Here. I originally built my AI workspace based on a Jeff Su video about cross-domain folder hierarchy. Shortly after, I watched Jake Van Clief's videos and the Interpretable Context Methodology (ICM). I thought what I was building was an ICM setup at first. But as I learned more, I realized I was doing something different. So I found a way to weld the two ideas together into a more comprehensive system. Here is how the integration works: Tiers 1-3 (The Reasoning System): This is the Jeff Su inspired structure. It handles routing and context delivery for different topics, sorted by Tier 2 domain. By the time you reach a Project at Tier 3, you hit the reasoning level, where the context and Claude md roles completely change depending on the work. These are organized by State/Kind rules embedded in the system, and reasoning sessions are distilled and indexed with key points synthesized to project context. Tier 4 (The Workflow): Once the reasoning is stable at T3, ICM comes in as the workflow layer underneath it. Figure out what you want to build in the reasoning Tier 3, configure the factory to build it in Tier 4. I've mapped out the basic folder structure in the diagram below, and attached the actual markdown doctrine file that runs it. CoworkOS Workspace Structure ============================ ROOT/ ├── CLAUDE.md (Tier 1: Root Map & Routing) ├── CONTEXT.md (Tier 1: Operator Contract & Global State) ├── 00_Resources/ (Tier 1: Global Reference docs) │ ├── Workstation_A/ (Tier 2: Broad Domain) │ ├── CLAUDE.md (Tier 2: Workstation Map & Routing) │ ├── CONTEXT.md (Tier 2: Domain Posture & State) │ │ │ └── Project_1/ (Tier 3: Reasoning Project) │ ├── CLAUDE.md (Tier 3: Project Identity) │ ├── CONTEXT.md (Tier 3: Live Reasoning Synthesis) │ ├── _reasoning-log/ (Tier 3: Spent thinking & ideation) │ ├── references/ (Tier 3: Settled constraints) │ │ │ └── tier-4-workflow/ (Tier 4: ICM Pipeline)

--------------------------------------------------------------------------------------------------------------------------------------- My name is Malcolm, I'm a recent Sound Engineering graduate out of Bristol U.K and in the last few months I've become obsessed with leveraging AI to make cool things. Ever since installing OpenClaw on an always-on laptop to the side of my desk and getting it to run a daily cron-job, I've been hooked! That was at the start of May, but I think I was jumping the gun a bit with this next part; looking for quick cash and convinced by AI gurus, I founded a humble AI-for-small-businesses and consulting company called 'Joy Automations' towards the end of the month to implement AI for small businesses in my area in Hampshire, England. Rightfully or not, I trusted that my few weeks of playing with OpenClaw would qualify me to earn top-dollar for implementing AI into local businesses, and I set out to speak to some about the services my new company could offer. I recieved mixed reactions and thinking from business owners about AI integrations, which is not unexpected for my beginner business: what was worrying me more was the complete inadequacy, incompetence, and forgetfulness that my personal agent would often exhibit when I'd ask it simple things. How was I to advice for local businesses if I couldn't even get my own agent to be reliable? That's when I found Jake's videos and ICM-related ideas, initially on social media. I didn't understand much of it at first, but I figured I'd pay attention since he's clearly set apart from many other AI creators on Instagram with his message, priorities, and style. Not trying to sell me something was a big part of that too. I have since watched his YouTube videos: Long story short, I've realized that it's worth it to slow down on the business side of things for now and get my foundations in order; to build a strudy agentic foundations for all of the work I do going forward, with ICM and durable context at the forefront!

used the methodology to code my first crazy idea, a digital "Jump to Conclusions Mat" from Office Space. I posted it to reddit, and the very first day I made .$40! This is probably the DUMBEST and most POINTLESS game on the internet, but hey, the guy who made pet rocks made a million dollars! I was so excited I can't even tell you! https://jumpoff.space/?utm_source=icm

I've been trying to figure out how to show someone what ICM actually looks like in under 30 seconds. You know how it goes — you start explaining numbered folders and stage contracts and human review surfaces, and their eyes glaze over around "stage contracts." So I built a demo that fits on a business card. Tap it to your phone, type what you do for work, and a live three-stage ICM pipeline runs in your browser. You watch each stage complete, read the output, pick what you want built, and walk away with a real workspace folder tailored to your work. All in about 20 seconds. ## The NFC Part I put an NFC chip (literally a small sticker) on the card. That's it. Each chip stores a short random ID — think license plate number. Tap the card, your phone reads the ID, sends it to a server, and the demo starts. I went with NFC over QR because tapping is just easier than framing with a camera. Also lets me tag each chip with context — I can attach "SF Tech Conference — dinner" on the backend and it shows up as a badge when someone taps. The URL carries nothing personal, just the random ID. The chip isn't the point though. It's just how someone triggers the demo. ## How the Pipeline Works Someone taps the card. Their browser opens to a text box: "What do you do for work?" They type something. Then three stages run, each feeding into the next: **Stage 1 — Discovery.** AI looks at what they typed and writes a quick analysis of their work and what parts are repetitive. **Stage 2 — Mapping.** AI designs a multi-step workflow for them. Each step has a note about what they'd want to review before moving forward. **Stage 3 — Opportunities.** AI suggests four specific projects they could build with ICM. Each stage streams back in real time. You can read what the AI produced before the next stage starts. That's the ICM thing — every output is visible and reviewable. Then they pick a project and the AI generates a real workspace. ## What They Walk Away With A folder of markdown files. Usually 5-7 of them. Numbered stages, instruction files per stage, a getting started guide. Download it, open it in Claude Code, type `setup`, and go.

There is a method of using Chinese models like: Qwen, MiniMax, Zai, Moonshot, etc. That are hosted in the US on the latest NVIDIA B300 datacenter GPUs, where your data stays private and is never trained on. This is where Ollama, the self hosted go to source for local models comes in. Ollama offers a selection of cloud based models. Simply append your model call with :cloud https://ollama.com/search?c=cloud&o=newest What's your favorite model?