Guide
Zettelkasten in 2026: why most people get it wrong
8 min · 2026-06-07
Niklas Luhmann wrote 70 books and 400 papers in 30 years. He attributed this to his Zettelkasten — a paper-card system of 90,000 atomic notes linked by reference numbers. When productivity culture rediscovered the method in 2020, it became a cult: Obsidian, Roam, Logseq, all promising to replicate Luhmann's output. Five years later, most users have a graph view they never open, and a folder of unfinished notes. This guide explains what Zettelkasten actually was, why translating it 1:1 to apps fails, and what to do instead.
What Zettelkasten actually was (not the romanticised version)
Luhmann's Zettelkasten was a research instrument for a single sociologist working on systems theory. Every card had a unique reference number that placed it in a branching sequence — not a tag, not a category, a position in an argument tree. Cards linked to other cards by their reference numbers, forming chains of thought he could follow weeks later.
The system worked because Luhmann had one continuous project: building systems theory. Every new card was already inside that project. The system was the project; the project was the system. Replicate that condition and Zettelkasten works. Don't replicate it, and you have an empty mansion of unlinked notes.
Why translating it to Obsidian / Roam fails for most people
Most people who try Zettelkasten in apps don't have one project. They have life: work captures, parenting thoughts, book quotes, half-ideas. The atomic-notes-with-links pattern only generates value when the notes belong to a coherent inquiry. Without that, you build a graph that looks beautiful, contains your life, and surfaces nothing useful.
The tools are blameless. Obsidian works. Roam works. The mismatch is between a method designed for one researcher's monoproject and a user living five overlapping lives. You either become Luhmann, or you let the system catch what you save and surface it when relevant — without the linking discipline.
What to do instead: capture, let the machine link, return when relevant
The Luhmann insight worth keeping is: ideas connect, and the connections are where understanding happens. The Luhmann mistake people copy is: humans must do the linking work manually. In 2026, AI can detect semantic connections in your captures with reasonable accuracy. You capture; the system links by meaning; the right note comes back when context matches.
This is the design of memory-layer apps like wamid. You don't tag, you don't link, you don't review. You drop captures in. The system reads them, finds semantic neighbours, and surfaces the right one when your current context matches. It's not Zettelkasten — it's the Luhmann insight delivered without the Luhmann labour.
Zettelkasten was a stroke of method genius for one person with one project. Translating it to your life means either becoming a monocentric researcher (rare) or doing busywork that produces a graph you never use. The Luhmann insight — ideas link, connections matter — survives without the bookkeeping. Capture, let the machine connect, trust the surface. That's what wamid is built for.