Blog

Record every thought

By Tristan Manchester·2026-01-28·Updated 2026-09-29·6 min read

Most productivity advice tells you to be selective: filter hard and only capture what matters. That assumes you know what matters when you have the thought.

You usually don't, which makes the filtering the problem. (For the loop I use instead, see Karpathy's append-and-review method, which Gravity is built on.)

Why filtering early fails

"Record everything" sounds undisciplined. Your brain produces a lot of noise, though, and random thoughts aren't a sign you're doing it wrong. They're samples. Collecting them is how you find the few valuable ones mixed in.

Filtering at capture feels disciplined, and it's usually premature. You're trying to separate signal from noise before you know what the signal looks like. The aim is to find the good thoughts later, when you have context, and you can't do that if you threw them away.

Quantity helps. Linus Pauling, one of only five people to win two Nobel Prizes and the only one to win two unshared, is usually credited with the line "The way to get good ideas is to get lots of ideas, and throw the bad ones away." Research on creativity points the same way: people who produce more ideas tend to produce more good ones, partly because you can't predict which attempt will be the standout.

The best ideas are rare and hard to spot in advance, and early filtering asks you to spot them in advance.

You also can't judge an idea well while you're having it. Generating ideas and evaluating them are different modes of thinking, often called divergent and convergent. The moment you ask "is this worth writing down?" you interrupt the thought, switch modes and make a permanent decision with almost no context. The time cost is tiny. The switching cost is bigger. Many of the thoughts you discard weren't bad, just unfinished.

Capture first, judge later

Writing something down is a form of processing as well as storage. Working memory is small (about three to five chunks), so every thought you hold while deciding its value takes room from the thinking you were doing. That's why shower ideas vanish before you reach your phone, and why "I'll remember this later" fails in meetings. The idea usually disappears because your attention moved on, whatever its quality.

Getting it out of your head frees that room and lets you look at your own thinking from the outside. That's why freewriting works: you write three throwaway lines, and the fourth shows you the real problem.

If "record everything" feels too vague, capture sparks rather than finished thoughts: questions, bits of friction, half-sentences, analogies, contradictions and obviously bad ideas. Openers like "I keep thinking about…", "It's odd that…" or "This is probably wrong, but…" are enough to get the thread down.

Let the mess decay, and review lightly

The obvious objection is that you'll drown in junk. You will if your system turns every old note into maintenance. Hoarding happens when each note keeps demanding filing decisions. If low-value notes can fade unless they earn fresh attention, most of the mess takes care of itself.

That's what sinking and rescue are for. Some thoughts deserve to disappear. If a note matters, it comes back: it catches your eye on review, connects to something you're working on, or keeps nagging. If not, it sinks and you don't have to feel bad about it. Treat it like compost: what looked like rubbish can feed something later.

An example. You write down "onboarding should feel like a game". On its own, it's weak. Three days later, support emails show people are confused during setup. On review the old note catches your eye, so you rescue it and add "game = progress feedback + small wins". A week later that line shapes the new onboarding. The value came from the collision, over time.

Review is where this works or fails, so keep it daily, quick and strict:

Open the stream and scroll fast.
Rescue anything that pulls at you.
If you like, add a line about why it mattered.
Stop after two minutes, even if there's more.

Two minutes keeps you noticing instead of cleaning up. A note earns a rescue if it jolts you, links to current work, repeats a theme, has a phrase you'll reuse, or would hurt to lose.

Why most notes apps make this hard

"Record everything" is bad advice in an app that punishes volume. In most notes apps each captured thought creates future work in folders, tags, titles and clean-ups. So people pre-filter to avoid the overhead, and lots of good but unfinished ideas never get written down.

Folders are predictions about where you'll look. Tags are bets that you'll remember your own vocabulary. Both add friction at the moment capture should be easiest. (I've written about why Gravity has no folders, and why prefixes are usually enough.)

Gravity changes the economics. Volume is cheap because sinking handles the clean-up. Write a thousand thoughts, let 995 sink, and rescue the five that keep proving useful. The app expects noise and gives you a way to find the signal in it.

So this isn't hoarding, because the decay is built in. It isn't journaling either: journaling is reflective writing, and this is a pipeline for ideas. Important notes don't vanish, because you keep rescuing them. And review stays light as long as you keep it short and focused on noticing.

Permission to be prolific

Record every thought, including the half-baked, obvious, embarrassing and unfinished ones.

Don't ask whether a thought is worth capturing when you have it. Write it down and let review and context decide what survives. In the moment, your only job is to get it out of your head.