Jisoo · February 22, 2026
Taste for Makers by Paul Graham
"...Copernicus's aesthetic revulsion toward [the equant] was one of the essential motives that led him to reject the Ptolemaic system..." - Thomas Kuhn, The Copernican Revolution
"All of us had been trained by Kelly Johnson, and fanatically believed in his conviction that a beautiful-looking airplane flies beautifully." - Ben Rich, Skunk Works
"Beauty is the first test. There is no permanent place in this world for ugly mathematics." - G. H. Hardy, A Mathematician's Apology
I recently spoke with a friend who teaches at MIT. His field is especially popular these days, so every year he is flooded with applications from students hoping to enter graduate school. He said:
"There are a lot of students who seem smart. But I am not sure they have taste."
Taste. We do not hear the word much these days. But whatever we call it, we still need the concept. By "taste," my friend did not mean merely a technically skilled student. He meant someone who could use technical knowledge to design something beautiful.
Mathematicians call good work "beautiful." Scientists, engineers, musicians, architects, designers, writers, and painters have used the same word, now and in the past. Is it just coincidence that they all use the same word? Or is there something common in the "beauty" they are talking about? If there is, might we be able to apply standards of beauty found in one field to another?
For those of us who design things, this is not merely a theoretical question. If "beauty" really exists, we should be able to recognize it. To make good things, we need good taste. Instead of treating beauty as a vague, abstract idea to be discussed at sentimental length or avoided altogether, let us approach it as a practical question.
How can we make good things?
These days, when you bring up taste, many people say:
"Taste is subjective, isn't it?"
It feels that way because it really does feel that way. People often do not know why they like something. It may be because it is beautiful, because their mother had the same thing, because they saw a movie star use it in a magazine, or simply because they know it is expensive. Many impulses and impressions are tangled together, but people do not examine them one by one.
Most of us have been taught since childhood not to analyze those tangled feelings too closely. For example, if you tease your younger sibling for coloring a person's face green in a coloring book, your mother will probably say:
"You do it your way, and they do it their way."
At that moment, your mother is not trying to teach an important truth about aesthetics. She just wants the two of you to stop fighting.
But these half-truths adults tell us often contradict one another. After years of insisting that taste is only a matter of personal preference, one day they take you to an art museum and say:
"Leonardo was a great artist, so look carefully."
What thought crosses the child's mind when they hear this?
What do they understand "great artist" to mean?
A child who has been told that everyone simply does things their own way is unlikely to immediately conclude that a "great artist" means someone who made better work than other people. In that child's worldview - a kind of Ptolemaic cosmology - a "great artist" is more likely to mean something that is good because a book says so, like broccoli: good for you, but for reasons you do not really understand.
Saying that taste is merely personal preference is a good way to avoid disputes. The problem is that it is not actually true. You feel this once you start designing things yourself.
Whatever people do, they naturally want to do it better. Soccer players want to win games. CEOs want to increase profits. Getting better at your work is a matter of pride, and it is genuinely enjoyable. But what if your work is designing things, and there is no such thing as beauty? Then there is no way to get better. If taste is only personal preference, everyone's taste is already perfect. You simply like what you like.
But as with any other job, if you keep designing, your skill improves. And your taste changes. Like anyone who gets better at their work, you know that you are improving. That means your earlier taste was not merely "different"; it was actually worse. At that moment, the axiom that there is no right or wrong in taste disappears.
Relativism is fashionable these days, and it can prevent people from thinking deeply about taste, even while their own taste is growing. But if you admit, at least to yourself, that good design and bad design exist, you can begin to study good design concretely.
How has my taste changed?
If I made a mistake, what caused it?
What have other people learned about design?
When you start examining these questions seriously, one thing is surprising: the ideas of "beauty" described by different fields have quite a lot in common. The principles of good design appear again and again across domains.
Good design is simple.
This point recurs in fields from mathematics to painting. In mathematics, it means that a shorter proof is generally a better proof. For axioms in particular, fewer is better. In programming, it means almost the same thing. In architecture and design, it means beauty should be based not on an excess of superficial ornament, but on a few carefully chosen structural elements. Ornament itself is not bad. It becomes a problem when it is camouflage for weak form.
The same is true in painting. A still life that carefully observes and solidly depicts a few objects is likely to be more interesting than a painting that repeats something like a lace collar in a flashy but meaningless way. In writing, the meaning is the same: say exactly what you want to say, and say it concisely.
It may seem strange that simplicity has to be emphasized. Shouldn't simplicity be the default? Flashiness actually requires more effort. But when people try to be creative, they fall into a strange state. A novice writer begins using a grandiose style nothing like the way they usually speak. A designer who wants to look artistic adds unnecessary curves and ornaments. A painter suddenly becomes an expressionist.
All of this is avoidance. Under the long words or the "expressive" brushstrokes, there may not be much there. And facing that is frightening.
When we are forced into simplicity, we confront the real problem. We can no longer hide behind decoration. At that point, we have to win on substance.
Good Design Is Timeless
In mathematics, every proof is timeless as long as it contains no mistakes. So what did Hardy mean when he said, "there is no permanent place for ugly mathematics"? He meant the same thing Kelly Johnson did: if something is ugly, it cannot be the best solution. A better solution exists somewhere, and eventually someone will find it.
Aiming for timelessness is also a way to force yourself to find the best answer. If you can imagine someone surpassing you someday, that means you should not leave the work to someone else; you should do it yourself. Some of the great masters did this so well that they left almost no room for those who came after them.
Every printmaker after Durer had to live in his shadow.
<Melencolia I (1514) - Albrecht Durer> Durer reached such an overwhelming level in printmaking, especially engraving and woodcut, that later printmakers could only be compared with his work. The pursuit of timelessness is also a way to escape the influence of fashion. Fashion, by definition, changes over time. If you can make something that will still look good in the distant future, its appeal is more likely to come from intrinsic value than from trend.
Interestingly, if you want to make something future generations will love, one way is to make something past generations would have loved too. It is hard to predict what the future will look like, but one thing is certain: the future will not care much about today's fashions either. So if you can make something that is attractive to people today and would also have been attractive to people in the year 1500, it is likely to work for people in 2500 as well.
Good Design Solves the Right Problem
A typical gas stove has four burners arranged in a square, with a dial for each one. So how should the dials be arranged? The simplest answer is to put them in a row. But that is a simple answer to the wrong question. Dials are used by people. If they are arranged in a line, the user has to pause every time to think about which dial controls which burner. It is better to arrange the dials in a square matching the burners.
A lot of bad design is diligent but pointed in the wrong direction. In the mid-20th century, it was fashionable to set body text in sans-serif type. Such typefaces are closer to the basic structure of letterforms. But that is not the problem body text has to solve. In readability, what matters more is whether letters can be easily distinguished from one another. Even if it looks somewhat Victorian, the lowercase g in Times Roman is easy to distinguish from the lowercase y.
The problem itself, not only the solution, can be improved. In software, difficult problems can often be transformed into different problems that are essentially the same but much easier to handle. Physics advanced faster when it stopped trying to reconcile natural phenomena with the Bible and instead turned to the problem of predicting observable phenomena.
Subscribe to Commits