🔄 Inversion of Control

The traditional model: the app owns the document. Word owns your .docx. Google owns your Doc. The app is the authority, the document is its captive.

The format serves the app, not the author. .docx doesn't exist to help you write — it exists to make your writing require Word. The format is the moat. The business model requires captive documents.

The pod was there before the app

A directory with files. A document and its assets, together in one place. That's the most natural thing in computing — it's how you'd organize stuff if nobody told you otherwise. It existed before apps, before GUIs, before platforms.

Then the suits said "no, the document lives inside the app." Because once the document is captive, the app is mandatory, and the app is the business model. The user is a customer, and they are already in the upgrade sales funnel as we speak.

Components are guests, not owners

In the pod model, the document is sovereign. Components inject themselves into the pod — the pod doesn't belong to the component. A terminal can appear inside a blog post. An oracle can appear inside a hexagram page. The document is the host, the component is the guest. Code is content.

Documents are stored on disk. Components are ephemeral processes with short lifetimes. The document was there before the component, and it'll be there after. Markdown and TOML — any text editor, any AI, any script can read and write them. The components are progressive enhancement. They show up if they're there. The content stands alone if they're not.

Why the suits will never build this

You can't monetize a directory. There's no lock-in in mkdir. The business model requires captive documents, and pods refuse to be captive.