Everyone knows compound interest. Fewer people notice that most knowledge work doesn’t compound at all.
You research a topic, write a deliverable, send it, move on. The research decays in your memory. The deliverable lives in someone’s inbox. Nothing feeds back. Next time you need similar research, you start from scratch — or worse, you half-remember something and don’t bother checking.
The problem isn’t effort. It’s legibility.
Legibility is the property of being readable and actionable by someone — or something — other than the person who created it. An insight in your head is illegible. Only you can access it, and only while you remember it. The same insight written down, structured, tagged, and linked is legible. Now it can be found, connected, and acted on without you in the loop.
This matters more than it used to.
When your only collaborator was future-you, illegible knowledge was merely inefficient. You’d lose some, rediscover some, muddle through. But when AI agents can search your knowledge store, score tasks against it, and execute work based on what they find — legibility becomes the difference between a system that compounds and one that just accumulates.
A messy folder of unsearchable notes is accumulation. A structured vault with consistent formatting, wikilinks, and metadata is capital.
The distinction isn’t about neatness. It’s about whether your past work reduces the cost of your future work. That only happens if something — your memory, a search engine, an AI agent — can find and use what you’ve already done.
Most people’s knowledge compounds in their head, which means it dies every time they context-switch. The structural move is externalising state into a form that agents can read and act on.
Not because you’re forgetful. Because compound interest requires a ledger.