A quiet argument for taste over output

This started as a throwaway line in a notebook and turned into the thing I get asked about most. So I will try to say it properly.

The version that becomes genuinely yours is the one where you kept the odd detail that did not come from anywhere else, the sentence that surprised even you. Protecting those small accidents is often where the real voice finally shows up, long after the imitation has been edited away. Originality, on the other hand, is mostly a willingness to leave strange things in. The first draft of almost anything is a collage of borrowed influences, and that is perfectly normal.

After a few years of this quiet refusal, you start making better choices faster, not because you have become smarter in any abstract sense, but because your standards have become detailed and muscular. You stop debating and start deciding, which is a much more useful posture to be in when the work is actually due. Taste is difficult to teach and easy to recognize, which makes it both oddly overrated and completely essential. It is the thing that tells you, before any of the data has arrived, that this version is the honest one and that other version is a compromise you will regret.

What I have come to believe is that taste is really just a name for a long history of caring. The person with good taste has simply paid attention more often, and to more subtle things, than the person without it. Developing taste is not about memorizing rules. It is about paying unusually close attention over a long period of time, until your preferences become specific enough to act on without thinking.

The patience is the whole practice, and the taste is just the residue it leaves behind. There is no shortcut to that. You build it the way you build any other muscle, through repetition and a willingness to be bored by your own early efforts until they stop being embarrassing. The good news is that taste compounds. Every time you reject something that is merely good enough, you calibrate yourself a little further.

If any of it is useful, take the part that fits and leave the rest. That is how most of this started for me anyway. The version that becomes genuinely yours is the one where you kept the odd detail that did not come from anywhere else, the sentence that surprised even you. Protecting those small accidents is often where the real voice finally shows up, long after the imitation has been edited away. Originality, on the other hand, is mostly a willingness to leave strange things in. The first draft of almost anything is a collage of borrowed influences, and that is perfectly normal. After a few years of this quiet refusal, you start making better choices faster, not because you have become smarter in any abstract sense, but because your standards have become detailed and muscular. You stop debating and start deciding, which is a much more useful posture to be in when the work is actually due. Taste is difficult to teach and easy to recognize, which makes it both oddly overrated and completely essential. It is the thing that tells you, before any of the data has arrived, that this version is the honest one and that other version is a compromise you will regret.