From Information to Creation

Information is cheap. What’s valuable is the ability to process it.

That sentence has, for the first time, become concrete in the AI era: a model can generate ten thousand words in a second, and it can bookmark, summarize, and retrieve almost anything for you — so “having information” has completely stopped being a moat. What matters now is “turning information into capability and finished work.”

Yet most people’s note-taking systems are still stuck at “dump everything in.” The bookmarks keep piling up, the knowledge base keeps growing, and the person behind it doesn’t change. The reason is simple: we treat information, records, knowledge, and creation as the same thing, when they are actually four completely different stages of work.

• Information is raw input — meant for AI to process or for you to skim quickly — and most of it is noise;
• Records are the semi-finished product between information and knowledge — the highest-conversion step, yet often mistaken for knowledge itself;
• Knowledge is structured, reusable capability sediment that solves your own problems;
• Creation is the finished product, recombined for an audience, that solves other people’s problems.

This column walks down that pipeline one stage at a time, with each essay solving the problem at one layer. The tools involved include Obsidian, Flomo, and Claude — the AI-era note-taking stack — and the cases come from a community of tens of thousands of members whose practices I’ve been following closely.

The column is ongoing — the overview is published, each of the four layers has its own essay, and more will follow.