Notes on Starting

The best time to plant a tree was twenty years ago. The second best time is now.

— Chinese proverb

The hardest part of almost any project I have ever worked on was not the middle, and it was not the end. It was the first thirty minutes — the part where nothing exists yet, and the full weight of the project is still an idea in your head.

Starting is hard for reasons that are usually psychological, not practical. The work itself is rarely the problem. The problem is that a project in your head is infinite and perfect, and the first moment you commit to a direction, you trade that infinity for a single finite thing. That trade always feels like a loss, even when the finite thing is objectively more valuable than the dream.

I have spent a long time thinking about this, mostly because I kept losing to it, and these are the notes I would give to a younger version of myself.

The First Draft is Supposed to Be Bad

The single most useful reframing I have ever come across is that the purpose of the first draft is not to be good. The purpose of the first draft is to exist. Once something exists, it can be edited. Before something exists, there is nothing to edit — there is only anxiety and unlimited possibility.

This applies to code, writing, design, product specs, and almost every other creative act. The blank page is doing something to your brain that the second page isn't. Once you have any text, any mockup, any prototype — even a terrible one — the shape of the work becomes visible, and your job switches from "invent something from nothing" to "improve this thing that exists." The second job is orders of magnitude easier than the first.

This is why professional writers talk about writing a terrible first draft as fast as possible. It is not a productivity hack. It is a way of sidestepping the part of your brain that is holding the project back.

Planning is Not Starting

The most convincing way to avoid starting is to plan.

Planning feels like work. It looks like work. It produces artefacts that resemble the output of work. But for most projects, planning past a certain point is a form of procrastination dressed up as diligence. You can plan a novel for a decade and still not have written a novel. You can spec a feature for six months and still not have shipped a feature. The plan is not the thing.

I am not arguing against planning. I am arguing for planning with a known stopping condition. The moment your plan is detailed enough that the first concrete step is obvious and unambiguous — stop planning. Take that step. You will learn more in the first day of actually doing the work than you would in another week of planning, and much of that learning will invalidate parts of the plan anyway.

The danger signal is when planning starts to feel comfortable. Planning should feel like warming up. If it feels like the project itself, you are substituting planning for starting, and the longer you do that, the more expensive starting becomes.

The Cost of Not Starting

We tend to think of not-starting as the safe option, because it involves no visible risk. This is an illusion. Not-starting is expensive, because the world is moving whether you are or not.

Every day a project sits unstarted, the gap between where you are and where you would have been if you had started gets wider. Your competitors ship. Your reference material goes stale. Your conviction about the original idea erodes. Your future self has to start from a harder position, against a smaller remaining window, than your present self would have.

The specific form this takes depends on the project. If it is a startup, the cost is that the market moves. If it is a book, the cost is that your taste drifts, and you may no longer be the person who can write that book in a year. If it is a personal project, the cost is usually that your life fills in around the empty space, and the space stops being available at all.

None of these costs show up on a balance sheet, which is why they are easy to ignore. But they are real, and they compound, and the hardest thing to do in your life five years from now is to have started today.

Small Commitments, Public Fast

The way I have found to actually start, reliably, is to make the smallest possible commitment I cannot back out of.

Concretely: write one sentence of the thing. Open a repo and commit the README. Send a short note to a friend describing what you are working on. The purpose is not to produce value — the purpose is to put a stake in the ground that your future self has to answer to. Once the repo exists, closing it takes an act. Once the friend knows you are working on it, abandoning it requires a conversation. These social and psychological costs are tiny, but they flip the default from "won't start" to "has started."

The reason this works is that starting is not a binary. Nobody goes from zero to fully committed in one jump. Starting is a sequence of escalating commitments, each small enough that the current you is willing to make it, but cumulatively enough that the future you finds it easier to continue than to stop.

If you get stuck, the correct move is almost never to plan harder. The correct move is to find the next smallest possible commitment and make it. One sentence. One commit. One sketch. One email. Whatever is too small to be scary.

Momentum, Not Discipline

Once you have started, the thing that keeps you going is not discipline. It is momentum.

Discipline is what the productivity books tell you matters. Discipline is what you need if you are trying to force yourself to work on something your brain does not want to work on. But for a project you actually care about, discipline is not the real fuel. Momentum is. And momentum is entirely a product of how recently you last touched the work.

A project you worked on yesterday is easy to pick up today. A project you last touched three weeks ago feels like a stranger. This is why every serious maker I know protects their streaks. Not because streaks are virtuous, but because a streak is momentum made visible, and momentum is the cheapest motivational state to maintain. Break it, and the restart cost is disproportionate.

The practical implication is that any amount of work on a project, every day, is worth more than a big burst of work followed by a week off. Five minutes of bad work keeps the momentum alive. A week of nothing kills it, and you pay for that a second time when you try to start again.