2. Resist the urge to link anything to anything

"Networked thought" is the new buzzword around information storage systems. While an app like Evernote gives you a rigid and pre-defined structure, apps like Notion and Roam let you link anything to anything. Instead of using folders and tags, people are building digital "zettlekastens," which are networks of notes that are interconnected through hyperlinks and meta-data. The idea of having an organic digital brain seems alluring, but does it actually improve our work?

These new information management tools are liberating, but also intimidating. Instead of just populating a pre-configured layout, you have to design and manage your own "information architecture." How should writers approach imposing order out of chaos?

The breadth of information we encounter is dizzying. When writers attempt to capture this frenzy, it can be overwhelming. This leads to the creation of systems that are "input" focused. Much time is spent linking notes to other notes. In many cases, writers will build structures that categorize inputs by medium (article, podcast, video), by length (indivisible ideas vs. long-form), by status (read vs. unread), or by originality (other's ideas vs. my own words). All of these strategies assume that having internal logic within our chaotic inputs will help our outputs.

In practice, designing and maintaining input-centric structures are one of the main things that can steal time from actually writing. If our top priority is to produce original work consistently, then we need to put librarian-like impulses aside, embrace a flat database of inputs with near-zero internal logic, and build an architecture that is output-centric.


Instead of building a structure that attempts to link and classify our source material, we should embrace chaos at the input level. A more useful tactic would be to ensure that each Input is linked to either an Output (an original thing I create), or a Cluster (a group of inputs that might one day feed into an output).

Our Inputs database can exist as a hot dumpster fire. Every input no matter it's attributes, can get dumped in here: books to read, book summaries, book ideas, podcast bullet points, shower thoughts, pictures of Tweets, Otter text, a meme, articles you read, articles you haven't read.

Instead of thinking, what other notes can an input link to, ask the question, "When will it be most useful to access this information in the future?" The decision to not grant myself permission to link notes to other notes, and ask this question instead, has caused a huge shift in momentum for me.

It's tempting in Notion to create as many databases as you can, that are each as specific as possible. That's how I started. But I learned the hard way that database minimalism save you so much headache. Our Outputs database can act as a single place to store all of our original work, regardless of the medium. We can use a parameter to differentiate essays, from newsletter drafts, from YouTube videos. If we had a database for each type of Input, then it creates our scheme of "relations" (how databases link together), more complex. There's also no need to create a separate database for essay ideas, when you can use a parameter to identify the status of a piece of content [backlog, on deck, in progress, draft, published].

If there is no clear Output for an Input, then you put it in a cluster. A cluster is a group of inputs centered around an idea. The idea is that from an Output template, you can easily pull in a Cluster to look for potentially related notes.