Capturing Everything, Understanding Nothing

In 2014, psychologists Pam Mueller and Daniel Oppenheimer ran an experiment at Princeton and UCLA that should have quietly settled most arguments about note-taking tools. Students who typed their lecture notes captured nearly three times as many words as students who wrote by hand. They also remembered significantly less. More capture, worse retention, consistently across multiple experiments. The mechanism wasn't mysterious. Transcribing verbatim doesn't require you to understand anything. You're a stenographer, not a thinker.

Most knowledge tools are optimizing for the wrong half of that tradeoff.

The collector's fallacy

Christian Tietze named it in 2014, the same year Mueller and Oppenheimer published their study, which seems fitting. He called it the "collector's fallacy": the mistake of treating the accumulation of information as equivalent to learning it. You bookmark the article, highlight the passage, clip the PDF, add the note to your system. Your brain registers this as task complete. The information is "in there." Except it isn't in you — it's in a folder somewhere.

Tietze was writing about Zettelkasten workflows, but the phenomenon is much older. Walter Benjamin, in his 1931 essay on book collecting, described the collector's relationship with his shelves as one of possession and memory. The books weren't primarily for reading; they were for having. The pleasure was in the acquiring, the arranging, the sense of completeness. We do the same thing with information now. The twelve-tab "read later" queue. The 800-item Instapaper backlog. The Notion database with 400 pages, none of them opened since Q3.

The act of saving something creates a feeling of security — this matters, I've preserved it — that substitutes for the harder work of actually processing it.

Why the digital version is worse

Physical hoarding at least has friction. Books are heavy. Paper piles up visibly. You notice when the system is broken because it occupies space you'd rather have back.

Digital capture has almost no friction, by design. Readwise syncs your Kindle highlights automatically. Apple Notes propagates across every device in milliseconds. Notion ingests entire web pages with a browser extension, title optional. This is called a feature. The research suggests otherwise.

A study published in Computers in Human Behavior in 2018 looked at digital hoarding behavior across 45 participants and found that the most consistent driver wasn't usefulness — it was anxiety. People saved digital files because not saving felt like loss, even when the saved items were never retrieved. The "just in case" impulse, applied to information, at infinite scale.

What gets lost in this loop is the one thing that makes information valuable: processing. Reading something slowly enough to disagree with it. Summarizing it in your own words. Noticing that it contradicts something you already believed. This is how information becomes memory, and how memory becomes judgment. None of that happens when you hit "save to Readwise."

What AI retrieval doesn't fix

AI tools are arriving with a plausible-sounding pitch: they'll retrieve things for you, so the organizational burden disappears. Capture everything; let the AI find the relevant piece later. The collector's fallacy solved by automation.

It doesn't work. If you haven't processed a piece of information, an AI surfacing it for you doesn't give you understanding. It gives you text. The AI can find the note you took three years ago, but it can't tell you whether the thinking in that note is still sound, or whether you've since changed your mind, or whether the context it was written in has dissolved. A database of unprocessed clippings isn't knowledge. It's a very organized kind of not-knowing.

The more interesting possibility is that structured capture changes the equation. Not "save this article" but "here is a preference," "here is a person," "here is a decision I made and the reason behind it." When you're forced to decide what something is before saving it, something different happens. You can't meaningfully add a preference record without thinking about what you actually prefer. You can't add a note about a person without asking what's true about them and worth keeping. The act of categorization is a small act of understanding.

Mueller and Oppenheimer found that handwritten notes were better because the physical constraint forced selectivity and paraphrase. Structure can do the same thing in a digital system: not as bureaucratic overhead, but as friction that earns its cost by requiring just enough thought to make the information actually stick.

The note that should never have been saved

There is a concept in library science called "the document that answers a question never asked." Most of what ends up in personal knowledge systems is exactly this. It felt important when you saved it. It answered a question you were asking then. You will probably never ask it again.

This isn't an argument against taking notes. It's an argument for taking fewer, better ones. For pausing before saving something to ask: am I capturing this because I've thought about it, or because saving it feels like doing something?

Usually it's the second one. Knowing that doesn't make it easier to stop.

---

Asgeir Albretsen is the founder of Harbor.