Attention is the only organisation system that works

Every note-taking system is trying to answer one question: how do you find the right note at the right time?

Folders answer it by location, tags by attribute, links by relationship and search by keyword.

Gravity answers it by putting the notes that recently caught your attention at the top and letting everything else sink. That sounds too simple to beat the clever schemes, and in my experience it beats them anyway.

The trouble with filing

Filing asks you to know how you'll want to find something before you know you'll want it.

When you put a note in a folder, you're predicting where you'll look for it. Work, Project X or Ideas? You're guessing. When you tag it, you're predicting what you'll search for. When you link it, you're predicting which other note it will matter to.

Sometimes you guess right. When you guess wrong, your future self can't see the note, because it's in the wrong folder, under the wrong tag or linked to the wrong cluster.

You also pay for every guess at capture time, on every note, whether you ever look for it again or not.

Attention doesn't have to guess

With attention-based organisation you decide what matters during review, when you meet the note again.

At that point you know what you're working on, who you've been talking to and what's on your mind. You're deciding with the most information you'll ever have, where filing makes you decide with the least.

If the note catches your eye, you rescue it. If it doesn't, it sinks. The order of the stream comes out of lots of small reactions, with no forecasting involved.

What "catching your eye" looks like

When you scroll your stream you don't read each note properly. You skim the first line and move on, and most notes pass without registering.

Some make you stop. You read the whole thing and feel something: interest, urgency, recognition. Psychologists call the underlying ability selective attention, your brain's knack for picking out what's relevant and ignoring the rest.

Usually the note stopped you because it touches something you're already thinking about: a current project, a conversation from yesterday, a worry you haven't resolved. You never linked the note to that project. It became relevant because your situation changed.

A folder can't notice that, and neither can a tag. You can, because you're the one who knows what your situation is today.

Sinking and archiving

In Gravity, a note you don't rescue sinks as new notes land above it. You don't have to decide anything. Leaving a note alone is the signal.

Archiving is a separate, deliberate step. When a to-do is done or a note has served its purpose, you swipe left and tap Archive (or press A on the Mac). It leaves the stream and goes to the Archive page, where you can still find it, and search can include it.

Between the two, the stream stays about the size of your current life. The notes you haven't touched in months are down at the bottom where they're out of the way, and the finished ones are in the archive.

Review as a sorting algorithm

Think of each review as a sorting pass. You start at the top and scroll down. Every note you pass is a quiet comparison: should this be higher than it is?

If you rescue something, you're saying yes, and it moves to the top. If you scroll past, you're saying it can stay where it is, and whatever you write next will push it a little further down.

After enough reviews, the notes that keep earning your attention collect near the top and the others drift down. You never sort anything on purpose. The sorting is a side effect of paying attention.

Doesn't the stream get overwhelming?

It's a fair worry. If everything starts at the top and only sinks through neglect, won't you end up scrolling through endless half-relevant notes?

In practice the inflow is small, a handful of notes a day for most people. Archiving takes finished things out. And because unrescued notes sink, your review only has to cover the top of the stream plus the occasional deeper scroll. You aren't re-reading every note you've ever written.

Most notes should sink. They did their job when you wrote them. The ones that stay near the top are the exceptions: ideas you're still working on, to-dos you haven't done, references you keep needing.

Doesn't it need constant review?

It needs regular review, which is less than it sounds. A minute or two of scrolling each day is plenty. If you skip a few days, the stream is still there, with a bit more to look at.

You don't need a weekly review ritual, inbox zero or maintenance sessions. Compare that with the upkeep other systems ask for: reorganising folders, pruning tags, fixing links, processing inboxes. Gravity's upkeep is scrolling and swiping.

What the top of the stream becomes

After a few weeks the top of your stream starts to show what you most recently cared about, which is often older than what you most recently wrote. It's the notes that survived review and that you keep rescuing.

It also corrects itself. When your focus changes, what you rescue changes, and the stream follows you. Nothing needed deciding upfront, and nothing needs maintaining.

If you want the thinking behind the method, Karpathy's append-and-review note is where it started. For why the sinking half matters as much as the rescuing, read the psychology of sinking.