Starting is exciting and everyone knows how to do it. The middle is hard but at least it has momentum. What almost nobody talks about is the last ten percent, where the interesting problems are solved, the ending is still slightly wrong, and no part of the remaining work will teach you anything new.

That flatness at the end is not a sign the project failed. It is what a nearly finished thing feels like from the inside. You have already learned everything it had to teach, so the remaining hours pay you nothing in insight. Expecting them to feel good is what makes people abandon work at ninety percent and start something new.

An abandoned project at ninety percent is worth less than a finished one at seventy.

Everything I have published has one detail I would still change. That is fine, and it might be necessary. The alternative is a folder of almost-finished things, each of them evidence of good taste and none of them read by anyone.

Finish it. Let it be a little wrong in public. Then use what you learn from the wrongness on the next one, which is a much better teacher than another week of private polishing.

On finishing things

At some point reading turned into a counting exercise. Books finished per year, pages per day, a shelf of covers as evidence of a life of the mind. I hit fifty-two one year and could not, in December, summarise a single argument from March.

A book is not a task with a completion state. Abandoning one at page sixty because it stopped teaching you anything is not failure, it is judgement. The only obligation you have to a book is to be honest about whether it is still doing something for you.

The books that changed how I think are, without exception, ones I have read at least twice. The first pass is orientation, learning the shape of the thing. Only on the second do you have enough of the whole in your head to see what the parts are doing.

Ten books read twice will change you more than a hundred read once.

Slow reading also means stopping mid-page to stare out of a window, which looks like nothing and is actually most of the value. The thinking happens in the pauses, not in the pages.

On finishing things

The boredom is structural

Ship it slightly imperfect