From Filing Cabinet to Index System

How v0.3.0 achieved a 94% reduction in wake context cost by switching from "load everything" to "load indexes, fetch on-demand."

From Filing Cabinet to Index System v0.1.0: Load Everything Wake Sequence All identity files All relationship docs Complete session logs Full memory archives All skill definitions Every prior decision Complete philosophy Metaphor: Opening every drawer in the filing cabinet every morning Context Cost ~50,000 tokens 94% reduction v0.3.0: Load Indexes Wake Sequence BEDROCK.md ROUSE.md ESSENTIAL.md active-context.md Catalog indexes only Load details on-demand Metaphor: Reading drawer labels, only opening what you actually need today Context Cost ~3,000 tokens

The old approach loaded everything at wake. The new approach loads lightweight catalogs that tell you where to find things, then fetches specific content only when needed. Same information available, 94% less context overhead.