On the "second brain" industry, gently

A field report from someone who has, over the past decade, tried every major "second brain" system and quietly returned to a plain notebook and a text file. What survived, and what did not.

I want to write about the "second brain" industry without disrespecting the people who work in it, because a lot of them are earnest and thoughtful, and one or two of them are close friends. But I have been an enthusiastic participant in that industry for close to a decade, and I have quietly walked back out of it, and I would like to explain why in a way that is useful to anyone doing the same walk.

The premise, restated fairly

The "second brain" idea, at its best, is a specific and defensible one: capture things you learn as you encounter them, connect them over time in ways your working memory cannot, and retrieve them later without the friction that killed most of your paper notes. The premise is right. The problem is not the premise. The problem is what happened to the premise once it became a product category.

Somewhere along the way, "keep a note when you learn something" became a professional-grade endeavour with a stack, a vocabulary and a set of best practices roughly the size of learning a second programming language. People I know spent — spend — hours a week grooming their systems. And when I look at the intellectual output of that time, I cannot honestly say it is bigger, or better, than the output of people who take casual notes into a plain text file.

What I tried, briefly

Everything. Zettelkasten, in three different apps. Graph-based note tools. Tag-heavy systems. Daily-note systems. Templates for meeting notes, book notes, project notes, retrospective notes. A short flirtation with a "PARA" hierarchy that I am still, occasionally, embarrassed about. In every case I started with real energy, produced a lot of well-organised notes for about six weeks, and then quietly stopped opening the app.

The system was elegant. It was also, in the end, a very expensive way to be alone with my notes.

The stopping was never dramatic. It was always the same shape: one week I did not review the graph, the next week the graph felt like homework, the third week I opened a fresh text file to jot something down "just for now," and I never went back. This pattern repeated across four different systems and about seven years. At some point I decided to take the pattern seriously.

What I do now

One text file called notes.md in a folder I keep open in my editor. New entries at the top. Each entry begins with an ISO date and, optionally, one or two hashtags. That is the whole system. When the file gets long enough that scrolling annoys me, I archive it into a folder called archive/, named notes-2025-Q4.md, and start a fresh one.

I search with grep. I do not link between notes. I do not have "atomic notes." I do not review anything on a schedule. When I need a note, I find it in about ten seconds because the search space is small and I remember roughly when I wrote it. When I don't need a note, I leave it alone.

What survived the migration

Three habits from the "second brain" era were genuinely worth keeping, and I still use them.

• Write it the moment you notice it. The gap between "I just realised something" and "I opened a file to write it down" is where 90 percent of notes go to die. Every system I tried was really a variation on shortening that gap. The plain text file shortens it further than any of them.
• Notes are for future me, not for anyone else. Every time I wrote a note thinking "maybe I'll turn this into a blog post someday," it was a worse note. The best notes I have are private, quick, and unpolished.
• Do not tag. Search. Tags are a fantasy of retrieval that grep replaces at a small fraction of the cost. Tagging is a promise you make to your future self about a taxonomy your future self will not, actually, respect.

What did not survive

• The daily note. I liked the idea. In practice, most days I did not have a thing to write in a daily note, and so I invented one, and the invented notes were noise.
• The graph view. Beautiful. Useless. In seven years of link-heavy note systems, I can count the number of times a graph traversal produced a genuinely useful connection on one hand.
• Template-driven meeting notes. My meeting notes got worse when they got templates. The template made me write in the shape of the template. When I dropped the template I wrote in the shape of the meeting, which turned out to be far more useful later.

A gentle offer to anyone in the middle of this

If you are, right now, in the middle of a rebuild of your "system" — pause. Open a plain text file. Write one week of notes into it, without any system at all. Do not tag. Do not link. Do not review. At the end of the week, look at what you wrote. Ask yourself, honestly, whether a more elaborate system would have produced more useful thinking or less. My guess is less, but you should test it on your own life.

Your mileage will vary. Some people genuinely do produce better work inside an elaborate system. But my own field report, after a long detour, is small: a plain text file, a bit of grep, and the discipline to actually write things down have out-performed every product I ever paid for.