This post is not about ICM, it's about learning the AI systems you are using in the ICM space. If you are new to ICM and you are looking to learn the structure and how to apply it, please see the posts below: (They will serve you well) 1️⃣Jakes Van Clief's Welcome: https://www.skool.com/cliefnotes/welcome-to-clief-notes-heres-where-to-start-2?p=f8f85a09 2️⃣Mira Bradsaw's collection of getting started: https://www.skool.com/cliefnotes/clief-notes-gems-for-starting-in-icm?p=0385cbf9 |---------------------------------Let's Begin---------------------------------------| 👇 📝Note: This post is not for people already working in the space, it's not for people who understand AI, it's for the people who are sitting in silence, watching the rest of us do those things. 🚦So, let's get started with real talk about what AI is here. A lot of smart people are quietly carrying around one perceived embarrassing question, what actually IS this "AI" thing everyone keeps talking about? If that is you, drop the embarrassment right now. You are not behind; we are all on the frontier of this thing called AI. The word AI has been stretched to mean everything and nothing, and there are many reasons for it, I will name the three reasons I see that create the most fog: 1️⃣People rushing to break into AI before they get it. 2️⃣People holding one piece of the puzzle like it is the whole thing, then screaming "I'm right!". 3️⃣People who just want to sell you something. If you notice a pattern, so did I and it's not the AI, it's the people pushing it! None of them are trying to make YOU capable. And that is why I love @Jake Van Clief and his team for creating a community which actually is! Today I start a journey to help remove the fog. 👇 The one-sentence version: AI is software that "LEARNED" to spot patterns from a huge pile of examples and uses them to make a good guess about something new. 💡

Hand a cheap model a task and it re-reads everything, or confabulates the one ID that actually mattered. So I pack it instead. A Context Pack is one file: exact needles (ids, paths, values) kept verbatim, the bulk gisted, sized to half the model's window. Workers read what matters and re-fetch the rest. A manager routes each task to the cheapest Claude tier that can do it, on your subscription, never metered. Small jobs go direct. Big ones fan out under a cheap large-context model that reassembles the result. I shipped this repo by pointing the tool at itself. It said "12 ok". I believed it only after diffing every file, because a workflow once told me "ok" for work it never ran. Verify the behaviour, not the tally. Repo, MIT, built with Claude: github.com/PUSHINGSQUARES/Build Deep-dive + a fidget to try: aris-space.com/documents/dispatch/context-pack-dispatch //A<3

I had no idea I wasn't actually using Claude Code this whole time, via the desktop interface... 😭 Attempting to give the desktop my 11labs voice agent to initiate workflows and discuss current operations within the build revealed the error.