4000 Cards Since 2019: What Zettelkasten Actually Looks Like in Practice
· 6 min read
Every few months someone discovers Zettelkasten and tweets about it like it's a new religion.
Atomic notes. Bidirectional links. Emergent ideas. The second brain that writes books for you.
I've been running the system since 2019. I have around 4000 cards. Let me tell you what it actually does.
What the Hype Gets Wrong
The Zettelkasten discourse, at least in English-language productivity circles, frames the system as a productivity accelerator. Write enough atomic notes, link them correctly, and insight will emerge automatically. The book writes itself.
This is not accurate.
The system is not an accelerator. It is a form of external memory with compounding characteristics. The compounding is real and substantial. But it takes years, not weeks, and the output is not automatic. You still have to write. The system just means that when you sit down to write, you are not starting from zero.
That distinction matters because the productivity framing creates unrealistic expectations. People start, maintain the system for three months, don't produce a book, conclude it doesn't work, and quit.
The Structure I Actually Use
I keep my cards in Obsidian. Plain text, locally stored. No sync service, no premium tier, no ecosystem lock-in.
Each card is one idea. Not one topic, one idea. The distinction: a topic is "pump cavitation." An idea is "cavitation in low-NPSH systems is often misdiagnosed as bearing wear because the vibration signatures overlap at certain RPMs."
One topic can have fifty ideas. One idea fits on one card.
Every card has: - A unique identifier (date + sequence) - One idea, stated in one to three sentences - Connections to related cards - A source or provenance note (where this came from)
That's it. No tags. No folders. The connections are the structure.
What 4000 Cards Actually Looks Like
I started in 2019. Slowly at first, a few cards a week, mostly from reading. Then it accelerated as I understood how to capture ideas from practice, not just from books.
Around 2021 I hit a threshold where searching the system started returning surprising results. I'd be writing something about documentation practices for Witanabe and find a card from 2019 connecting it to something I'd read about archival standards. I had no memory of making that connection deliberately. The system had it.
This is the compounding that the hype mentions but doesn't explain well. The cards are not just storage. They are a record of how you were thinking at different points in time. When you connect a 2019 card to a 2022 card, you are creating a synthesis that neither version of you could have written alone. The earlier you didn't have the experience. The later you had forgotten the conceptual clarity of the earlier reading.
At 4000 cards, I have material for somewhere around 40 essays that I haven't written yet, because I can see the connection clusters but haven't had time to write them up. That backlog is not a failure of the system. It is evidence that the system is working. Ideas are accumulating faster than I can publish them. That is a good problem.
Where the Cards Come From
About half the cards come from reading. Books, papers, documentation. I read slowly and annotate. Then I sit with the annotations and extract the ideas worth keeping. Not summaries, ideas. "The author's main point" is not a card. "The reason X fails in context Y is Z, and the implication for my situation is W" is a card.
The other half come from practice. This took me longer to figure out.
When you solve a problem in the field, there are usually one or two things that surprise you. A pattern you didn't expect. A failure mode that contradicts what the manual says. An approach that worked for reasons you can't fully articulate yet. These are the most valuable cards because they are the ones nobody else has. They are not available in any book.
The discipline is to write them down immediately or close to it. Not in polished form. In the form of the idea as it appears to you in that moment. You can refine later.
The Failure Mode Nobody Talks About
There is a failure mode in Zettelkasten that is more common than "the system doesn't work."
It's that the system works as storage but not as a writing engine, because the cards are too granular and too disconnected from the projects you actually want to produce.
This happened to me in 2020. I had 800 cards and couldn't figure out how to turn them into anything. Every card felt like it needed five more cards before it was complete enough to use.
The fix was to add what I call project cards. Not atomic ideas, but working notes for a specific piece I'm trying to write. The project card pulls in relevant atomic cards and gives me a place to assemble them into an argument. The atomic cards stay unchanged. The project card is the assembly workspace.
This is not in most Zettelkasten tutorials because it violates the "everything must be atomic" rule. But in practice, you need a layer between the atomic idea and the finished piece.
What It Has Actually Produced
The 558-title publishing catalog that Hibrkraft maintains didn't start from scratch. It started from notes. Not Zettelkasten in its current form, but the same underlying habit: when you have a thought worth keeping, write it down. When you have enough written thoughts on a topic, you have a draft.
The essays on this site come from the system. Not written by the system. But they start as card clusters, not blank pages.
The methodology documentation for Witanabe projects draws from cards accumulated over five years of working in industrial engineering. I can reconstruct why certain decisions were made because I wrote down the reasoning at the time, not afterward.
This is the practical value: not emergence, not automatic synthesis, not the book that writes itself. The value is that your past thinking becomes available to your present thinking in a structured way. You get to stand on your own shoulders.
Starting From Zero
If you are starting now:
Write one card per idea per reading session. Not per book. Per reading session. You will not miss important ideas by being selective. You will miss them by trying to capture everything and drowning in volume.
Connect cards when a connection is obvious. Don't force connections to make the system look smart.
Start using the system for something real within the first month. A note, an email, a short post. The habit of retrieval is as important as the habit of capture.
Expect the compounding to become visible around the 500-card mark. Before that, it will feel like expensive note-taking.
After that, it starts to feel like thinking with more of yourself than you can hold in your head at once.
That is what it actually does. Nothing more, nothing less.