Working in the Open

A long essay, written partly to think and partly to test how this site holds up when a page runs well past a single screen. Scroll, and the outline on the right should keep pace.

Most of what I make, I now make in public. Not because it is brave, and not because it is good for some imagined audience, but because the alternative — working in a sealed room and emerging only when finished — quietly corrodes the work itself. This is an attempt to say why, carefully, without turning it into a manifesto.

The closed room

There is a particular fantasy that surrounds private work. You picture yourself disappearing for six months, emerging with something whole, and presenting it to an astonished world. The fantasy is seductive because it flatters the ego: it casts you as the lone author of a finished thing.

The reality is less flattering. Closed rooms accumulate dust. Without anyone looking, the work drifts toward whatever is easiest for you specifically, and the gaps between what you intended and what you built go unnoticed until they are expensive.

What secrecy protects

Secrecy protects two things, and it is worth being honest that both are real. It protects unformed ideas from premature judgement, and it protects your sense of authorship. An idea spoken too early can be argued out of existence before it has had a chance to grow its own defenses.

So I am not arguing that everything should be visible all the time. The seed stage deserves a closed room. The argument is about what happens after the seed — when the thing has a shape but not yet a finish.

What secrecy costs

The cost is feedback, and feedback is the only thing that reliably improves work. In private, you are the entire quality function. You decide what is good, you decide what is done, and you decide when to stop — and you are, on all three counts, the least reliable judge available.

The work you are too close to is the work you can no longer see.

The closed room also costs you momentum. Nothing external pulls the work forward; there is no one waiting, no one to disappoint, no small public embarrassment to avoid. Left alone, projects expand to fill all available secrecy.

The open bench

The alternative I keep returning to is what I think of as the open bench: a visible surface where the half-finished thing sits, and where anyone who cares to look can see exactly where it is.

This is not the same as performing. Performing is showing the highlight reel. The open bench is showing the actual state, including the parts that are ugly and the parts that do not work yet.

Visibility changes the maker

The first thing the open bench does is change you. When the work is visible, you hold yourself to a standard you would not bother with in private. You write the comment. You name the variable properly. You fix the thing that only you would ever have seen.

This is not vanity, or not only vanity. It is that visibility recruits a part of your attention that private work leaves idle. You think more clearly when you suspect you will have to explain yourself.

Visibility changes the work

The second thing is slower and more important. Over months, visible work attracts correction. Someone notices the flaw you had stopped seeing. Someone asks the question you had been carefully not asking. The work gets better not because you are smarter but because the room is bigger.

There is a failure mode here, which is mistaking attention for correction. A large audience that only applauds is no better than an empty room. What you want is a small number of people who will tell you when you are wrong.

How to actually do it

Principles are cheap. The harder question is what working in the open looks like on a Tuesday, when you have ninety minutes and a half-broken idea.

Lower the stakes of publishing

The single biggest barrier is that publishing feels like a verdict. If hitting "publish" means "I am declaring this finished and good," you will almost never do it. So the trick is to make publishing mean something smaller.

• A note is not an essay. It can be three sentences and a question.
• A draft is allowed to be wrong, as long as it is marked as a draft.
• A log entry records what you did, not what you achieved.

When publishing costs almost nothing, you do it often, and frequency is the whole game. The compounding only starts once the activation energy drops near zero.

Separate the seed from the bench

Keep a private space for seeds — the truly unformed, the things that would die if spoken aloud too soon. Then move things to the open bench deliberately, when they have enough shape to survive contact with other people.

The mistake is to have only one of these. All-private, and you lose feedback. All-public, and you lose the protected space where fragile ideas can form. The boundary between them is where the judgement lives.

Write the boring parts down

Most of working in the open is unglamorous documentation: what you tried, what broke, what you decided and why. This is tedious in the moment and invaluable later. The reasoning behind a decision evaporates within days; the decision itself remains, now mysterious.

2026-03-04  Tried the single-cycling toggle instead of three buttons.
            Reason: the corner UI was getting crowded; one control that
            previews its next state reads cleaner. Revisit if users miss
            direct access to "reader mode".

A log like this is the cheapest possible insurance against your future self, who will remember none of this and will otherwise be condemned to relitigate every settled question.

What it does to fear

Underneath all of this is fear, and it would be dishonest to pretend otherwise. The fear of working in the open is the fear of being seen to be wrong, or unfinished, or ordinary.

The audience is smaller than you think

The first consolation is statistical: almost no one is watching, almost all of the time. The imagined crowd — poised to judge your every misstep — does not exist. What exists is a handful of people, most of them busy, a few of them kind, and the occasional stranger who found the thing useful.

Being wrong in public is a skill

The second consolation is that being wrong in public is a skill, and like any skill it dulls the fear through repetition. The first time you publish something flawed and survive, you learn that the catastrophe was imaginary. The tenth time, you barely notice.

This is, in the end, the quiet benefit that outlasts all the others. Not better work, though the work does get better. Not feedback, though the feedback is real. It is the slow erosion of the particular fear that keeps so much good work trapped in closed rooms, unseen and unfinished, waiting for a confidence that only visibility could ever have provided.

A short coda

If you have read this far, you have scrolled a good distance, which means the outline on the right has had something to do. That small mechanical fact — a list of sections quietly tracking where you are — is itself a tiny argument for the open bench. It only works because the structure of the thing is visible.

Make the structure visible, and a great deal else follows.