Six years ago, in the very first piece I wrote here, I described the small magic in our pockets. The magic that stretches memory almost forever, and blurs the line around who owns knowledge. Since then my notes kept growing. The magic kept its promise and pushed my memory further and further outward.

And now I stand on its other side. Too many to find.

Organizing had become the work

If you use Obsidian, you probably know the feeling. Past a few hundred notes, you can no longer reach the one you know you wrote. All that remains is the déjà vu — “I’ve written this before” — while the note itself stays hidden.

So you organize. You recut the folders. You settle on a tag system. You relink. It works, at first. But as the pile grows you end up back in the same place. Before long you spend more time organizing than writing. Organizing itself had become the work.

Recut folders, added tags, and it collapsed again

Everything I tried was an attempt to solve it by place. First the folder hierarchy. Then tags. When that wasn’t enough, maps of content tied together by links. I even tried Bases. I wove ontology-style cross-links between notes too, of course — and still, things slipped through and went uncaught.

Each one feels good the moment you cut it — the shelves line up, the world seems legible. But half a year later the shelves overflow again. I can’t remember which drawer I put it in. The more places you add, the higher the cost of remembering those places. It was never a real fix.

There were two kinds of organizing

At some point I noticed that the single word “organizing” was really two piles of a different nature.

One is hunting for duplicates and relations by hand. Tracing memory, opening folder after folder, just to check “did I write this already?” This grows heavier in proportion to the pile. When a human does it, it only drains you and produces nothing. It is, by rights, the machine’s work.

The other is facing the text itself. Merging, splitting, renaming, deciding what to keep and what to throw away. In code we would call it refactoring. This is the body of intellectual work — the good kind, the kind I want more time for.

My mistake was not separating the two. The drain of the former eats the time of the latter. The good work gets pushed out by the bad.

A question that place cannot answer

The decisive moment came when I went looking for a note. What I wanted to recall was this:

“What became important in my career, and what I learned from failing at it.”

You cannot pull that with grep, or folders, or tags. There is no folder called “what became important.” You could never have tagged it in advance. People remember by meaning, not by where they filed it. The very premise of searching by place no longer fit.

Discovery you may hand off. Judgment you keep.

From here it stops being about which tool to use, and becomes about where to draw the line.

Discovery you may hand off — the work of finding duplicates and pulling the near-in-meaning ones toward you. In fact you should. The more you do it by hand, the more it drains you and loses to sheer volume.

But judgment you must not hand off. What is the same and what is different. What to keep and what to discard. That understanding stays on your side. It is the “understanding” side of a piece I once wrote, We Must Not Delegate Understanding. Understanding means connecting the dots yourself. The moment you hand that to something else, your own mind quietly goes shallow.

This has long been ordinary in software. No one hunts for duplicate code by eye. You leave it to the linter, to “find usages.” The human moves only on what comes after — the judgment of how to rewrite. Notes should have been the same.

After that

Once I made peace with this line, I began building a small mechanism to return the “discovery” side to the machine. Nothing is sent to someone else’s server; it stays at hand, and I hold the key.

I haven’t stopped facing my notes. I’m only letting go of the part a machine can do. The time it frees I return to facing the text — merging, splitting, throwing away, and writing again.

Next I want to write about how to keep doing that judgment. Because the assurance that a future me can surely find it again is what makes forgetting a happiness.