Lately, besides creating my own original content, I’ve also been browsing a ton of information online every day. Among them, the one I most want to compile today is Paul Graham’s “Taste for Makers.” Actually, I read this article a long time ago, but I think it’s worth revisiting and sharing.

Paul Graham: What Makes Good Design?

Original author: Paul Graham | Revised and compiled in Chinese translation

When building a product, writing an essay, or designing an everyday object, we eventually all face the same question: how do we judge whether it’s any good?

“I like it” is certainly one answer. But when you need to build something yourself, knowing your own preferences is often not enough. You have to see where things fall short, understand why they fall short, and figure out how to improve them.

In Taste for Makers, Paul Graham discusses exactly this kind of judgment. The “taste” he talks about isn’t just about what you appreciate — it’s about whether you can recognize quality and make your work better.

This essay touches on painting, architecture, mathematics, and software. Each field has its own distinct problems, yet certain qualities in great work keep appearing: simplicity, endurance over time, solving the right problem, and a willingness to revise again and again.

They aren’t a checklist you can mechanically apply, but they can help turn a vague sense of “this feels right” into something more concrete and actionable.

Taste Can Improve

People often say taste is subjective. You like one color, I like another — there’s no need to argue over which is better.

This statement has its uses. Respecting personal preference saves us from pointless arguments. The problem is that if every judgment about quality is reduced to preference, “making it better” loses all meaning.

Once you start creating, you’ll find your judgment shifting. Writing that once felt fluent may now seem sluggish; interfaces you were once satisfied with may reveal unnecessary steps. You haven’t just changed your preferences — you may have genuinely understood problems you couldn’t see before.

This is why it’s important to distinguish personal liking from judgments about quality. We might like something because of childhood memories, or because it’s expensive, popular, or used by someone famous. These factors are real, but they don’t directly tell us whether a design is good or bad.

Likewise, acknowledging that designs can be better or worse doesn’t mean everyone’s taste must converge, nor does it mean every artwork can be judged like an arithmetic answer. Graham’s argument is better understood as a working attitude: a willingness to question your own preferences, study the work, and admit there’s still room for improvement in your judgment.

For creators, this attitude is practical. It saves us from hiding behind “that’s just my style” and lets us keep asking: what’s still not good enough? And why?

01 Good Design Is Simple

From mathematical proofs to paintings, from programs to architecture, simplicity is a recurring hallmark of excellent work.

But simplicity doesn’t mean fewer elements, and it doesn’t mean turning everything into blank space. It means knowing which parts do real work and which just make the piece look more like a finished work.

An article stuffed with long sentences and jargon may just be one where the thinking isn’t clear. An interface full of animations and decoration may just be one lacking solid structure. Complexity on the surface sometimes covers for weakness underneath.

A useful test: strip away the decoration — does the work still hold up? Is the underlying logic clear? Is the structure sound? Is the core function worth using?

Decoration itself isn’t the problem. The question is whether it enriches the work or masks its hollowness. Simplicity forces us to confront this.

Fitz Henry Lane, Becalmed off Halfway Rock, 1860. An illustration for observing how a few subjects, a horizontal structure, and large areas of negative space work together to organize a composition.

02 Good Design Endures

Trends can bring something attention quickly, but they don’t guarantee it will hold up over time.

If a design’s appeal comes mainly from “that’s what everyone’s doing now,” it’s just as easy to lose its appeal when the trend shifts. Design that endures relies more on its own merits: proportion, structure, function, and the accuracy of its expression.

Ask yourself: when today’s fashionable colors, fonts, and visual motifs fall out of favor, what’s left that’s worth keeping?

This isn’t to say all work should chase some timeless style. Different eras have different materials, technologies, and ways of using things. What’s worth guarding against is letting trends replace judgment — assuming that keeping up with the latest style is the same as finding a good answer.

Viewing your work on a longer timescale helps redirect attention back to the problem itself.

03 Good Design Solves the Right Problem

A design can be tidy, beautiful, and labor-intensive — and still solve the wrong problem.

Graham gives the example of a stove. Four burners arranged in a square, but the control knobs lined up in a row. For the designer, a row of knobs looks neat. For the user, every operation requires figuring out which knob controls which burner. If the knobs mirrored the layout of the burners, the correspondence would be immediately intuitive.

The key here is the user’s process. What the designer needs to solve is enabling the user to complete their task smoothly — not just making the control panel look tidy.

Software and tools work the same way. No matter how polished the interface, if users have to double-check, take detours, or deal with things they never needed to deal with, the design hasn’t done its job.

Sometimes improving a design starts with reframing the problem. First understand what people are actually trying to accomplish, then decide which features, structures, and visual arrangements will help them get there.

Spatial correspondence reduces cognitive cost. The numbers in this diagram illustrate the mapping only — they don’t represent labels a real product must have.

04 Good Design Leaves Room

Good work doesn’t have to spell everything out, nor does it have to arrange every possible action for the user.

Jane Austen’s novels are one example cited in the essay. She doesn’t rely on extensive scene-setting; instead, she lets readers imagine the world around the characters from the story itself. The right clues involve people far more effectively than exhaustive description.

Design works similarly. A tool can offer clear core capabilities that users combine as they see fit. A building can support different ways of living, without requiring everyone to live the way the designer envisioned.

This openness isn’t vagueness. The clearer the fundamental rules, the easier it is for users to explore. Missing essential information only creates confusion.

Good design explains what needs explaining — and leaves room for people to discover and create on their own.

05 Good Design Has a Sense of Humor

Here, humor doesn’t mean deliberately inserting jokes, nor making everything comical.

It’s closer to a kind of ease: the work has strength, but doesn’t need to strike a serious pose at every moment. When the creator is in command of their expression, there’s room to stay light — even playful.

Graham links this sense of humor to confidence. But he doesn’t present it as a hard requirement that applies to every piece of work.

So it would be wrong to conclude that serious work must be lesser, or that a lack of humor means the creator lacks control. The more useful takeaway is this: taking your work seriously doesn’t have to mean taking yourself too seriously.

06 Good Design Requires Hard Work

Excellent work often demands enormous time and effort. A clean interface, a comfortable chair, a smooth-reading article — behind each are often many unseen attempts.

Truly difficult problems force creators to make choices. When resources are scarce, every element has to justify its necessity; when conditions are harsh, flashy but useless parts are unlikely to survive.

But difficulty itself isn’t something to romanticize. Some difficulty comes from the depth of the problem; some comes from messy collaboration, shifting requirements, or unreliable conditions. The former can push design forward; the latter usually just wears people down.

So this principle shouldn’t be read as “the more suffering, the better the work,” nor should it be used to dismiss light, enjoyable working conditions. What’s worth investing in is the effort the problem actually demands — not suffering endured to prove you tried hard.

07 Good Design Looks Effortless

Elite athletes make difficult movements look natural, and great design often has this quality. People see the finished product and feel it simply had to be this way.

But this naturalness usually doesn’t mean the process was easy. An article that reads like casual conversation may have been rewritten many times. A tool that feels effortless to use may have gone through extensive adjustment.

It’s the complex judgments left behind in the process that allow the final form to come out clear and precise.

This is related to simplicity, but not identical. Simplicity is about a work’s structure; looking effortless is about its fluency in use and experience.

Nor is there any need to deliberately hide every trace of the process. Textures left by hand, visible brushstrokes — these can all become part of the work. What matters is whether they serve the work, rather than asking the viewer to pay for the creator’s toil.

08 Good Design Makes Use of Symmetry

Symmetry creates order and makes complex things easier to understand. It can appear as repetition, or as recursion: a structure that reappears within its smaller parts.

Repetitive rhythm can organize buildings, music, and essays; smaller structures can combine to form larger ones. Designers don’t need to invent a new set of rules for every local detail.

Symmetry here isn’t limited to mirror-image arrangements, but it doesn’t reject geometric symmetry either. A more accurate way to put it: design can employ symmetry at different levels, without requiring the whole work to be uniformly left-right symmetrical from end to end.

Symmetry also carries risks. When repetition is merely filling space or applying a ready-made tidy form, it can substitute for thinking.

Order is worth pursuing, but one question still needs answering: what does this order actually serve?

09 Good Design Learns from Nature

Nature offers rich reference points for design. Form, structure, and modes of movement can all inspire us to rethink the problems at hand.

Learning from nature doesn’t mean dressing objects in leaves or animal shapes. It asks us to observe: how does a structure bear load? Why does a particular form suit its environment? Which relationships can be carried over into a new design?

Likewise, looking similar doesn’t guarantee the solution will work. Materials, scale, forces, and conditions of use differ — a borrowed structure may need substantial adaptation.

So it’s misleading to describe nature as a system that automatically eliminates all redundancy and outputs perfect answers. What’s more valuable to the designer is the patience to observe and understand, treating nature as something to learn from.

10 Good Design Is Revised

Few people get things right on the first try. Mature creators leave room for revision and are willing to admit that some of their earlier work may need to be redone.

The difficulty is often psychological, not technical. Having already invested a lot of time, you’re reluctant to scrap it; if a part turned out well by accident, you worry redoing it will make it worse. So it’s easy to find reasons to justify the result in front of you and convince yourself it’s good enough.

But a healthy degree of dissatisfaction with your own work is what drives continued improvement. Being able to see problems — and willing to deal with them — is what creates the chance to move the work forward.

Writing needs editing. Code needs refactoring. Products need adjustment through real use. But a first draft isn’t necessarily worthless, and starting over isn’t the only path to progress. Knowing what to keep and what to fix is itself part of design ability.

Making errors easy to spot and revisions easy to make is often more useful than trying to maintain the image of getting it right the first time.

Revision isn’t redoing everything from scratch. It’s identifying problems, making choices, and checking the results.

11 Good Design Borrows

Creators don’t need to avoid every good solution that came before, just to prove their originality.

But borrowing operates at different levels. Copying form alone doesn’t mean you’ve understood the work. Knowing where a method comes from, why it works, and under what conditions it holds — that’s what allows you to apply it to your own problem.

Pursuing originality has value. The problem is when “being different from everyone else” becomes the sole goal — it easily leads to pursuing difference for its own sake, or even abandoning approaches that have already proven effective.

A more mature attitude puts the quality of the work first. Others’ answers can serve as a starting point, but you still need to judge whether they fit your current situation, and then do your own work.

That’s why “borrowing” is a more accurate term than “plagiarism.” This principle is about how to learn from and apply existing experience — it provides no justification for taking credit for someone else’s work.

12 Good Design Sometimes Looks Strangely

Some great works feel unfamiliar — even carrying an unexplainable beauty. But “strangeness” is not a sufficient condition for greatness, and certainly not a shortcut to pursue directly.

If the creator is only thinking about being different, it’s easy to mistake a strange shape for an achievement without actually solving the problem.

A more reliable path is to focus on doing the work well, and let style emerge gradually from your choices, judgments, and sustained effort. Everyone notices different details and handles problems differently — these differences naturally find their way into the work.

So there’s no need to put “developing a personal style” at the front of every project. Let the work stand on its own first; and when something distinctive appears, don’t rush to press it back into a familiar shape.

13 Good Design Often Appears in Clusters

Great work tends to appear in concentrated bursts — certain periods, certain places. Renaissance Florence is one example cited in the essay.

This reminds us that creation is not purely the result of individual talent. What the people around you are studying, what works keep coming into view, and what standards the community holds for quality — all of these shape a person’s judgment and growth.

A good environment helps you spot your own weaknesses faster and gives you access to problems worth pursuing. Discussion, comparison, and collaboration among peers continuously push the work forward.

But this doesn’t mean you must find a “genius cluster” before you can start, or that simply joining the right circle will automatically produce good work.

A more practical approach: seek out people doing serious work, keep yourself surrounded by excellent results, and put yourself in conversations where you can get concrete feedback. The environment provides opportunity — the work is still yours to do.

14 Good Design Often Requires Courage

In every era and every field, there are practices protected by habit. Because everyone has always done it this way, they look like answers that no longer need discussing.

Improvement sometimes begins with an uneasy feeling: this doesn’t make sense, this process could be smoother, there might be a better possibility for this form.

Turning that feeling into work requires courage — and understanding. Merely opposing convention doesn’t prove you’re right. You need to understand why the convention exists before you can judge whether it’s still necessary.

The value of boldness lies in the willingness to question what’s taken for granted, and to take on the work of finding a better answer.

If a change can only be explained by “other people just don’t get it,” that’s a reason to stop and look again. Good design needs to challenge old judgments — and be open to new scrutiny as well.

Bringing Judgment Into Practice

After reading these principles, there’s no need to rush and sort every piece of work into fourteen categories for scoring.

Start with something in front of you: an article, a page of interface, a table. What do you think is good about it, and what seems off? Does this judgment come from personal memory, from trends, or from observing the structure and the process of use? If you were to improve it, what would you change first?

These questions make appreciation more concrete — and keep creation from relying entirely on intuition.

Taste can be cultivated gradually through observation, comparison, and hands-on work. Acknowledging that your work isn’t good enough yet isn’t a rejection of creativity — it’s a direction for the next round of revision.

You only truly understand how much you’ve grasped once you put your judgment to work on an actual piece.

Original source: Paul Graham, “Taste for Makers,” February 2002. https://paulgraham.com/taste.html

Statements checked against the original English essay; the fourteen principles are preserved. This is a revised compilation, not a line-by-line translation. Rights to the original text and its original compilation belong to their respective owners.

Cover photo: Alex Lewin, from Paul Graham’s official images page (https://www.paulgraham.com/images.html), noted there as public domain. The cover uses only proportional layout and text typesetting; the figure has not been generated or altered.

Illustration in article body: Fitz Henry Lane, Becalmed off Halfway Rock, 1860, National Gallery of Art; image from Wikimedia Commons, CC0. The two diagrams were created specifically for this compilation. None of the illustrations here were included in Paul Graham’s original essay.