| name | taste |
| description | Evaluate and improve the taste and design quality of code, architecture, APIs, UI, writing, or any creative work using Paul Graham's principles from 'Taste for Makers.' Use when the user asks to review something for taste, elegance, beauty, or design quality. Also use when the user says 'taste check,' 'taste review,' 'is this elegant,' 'make this beautiful,' 'review for taste,' 'does this have good taste,' or wants to improve the aesthetic quality of their work beyond mere correctness. |
Taste for Makers
Taste is not subjective. As you get better at designing things, your tastes change — and you know they're getting better, which means your old tastes were not merely different but worse. Good and bad design exist. This skill helps you see the difference.
The recipe for great work: very exacting taste, plus the ability to gratify it.
It is easier to see ugliness than to imagine beauty. When reviewing work, listen for the voice that says "What a hack! There must be a better way." Cultivate that voice. Don't ignore it.
The 14 Principles
1. Simple
Shorter proof, fewer axioms, less ornament. When forced to be simple, you face the real problem. Ornament is camouflage on insipid form. When you can't deliver ornament, you have to deliver substance.
Ask: Can anything be removed without loss? Is complexity here earning its keep, or hiding a lack of substance?
2. Timeless
If something is ugly, it can't be the best solution — a better one exists. Aiming at timelessness evades the grip of fashion. If it would have appealed in 1500 and appeals today, it will likely appeal in 2500.
Ask: Will this still look right in 10 years? Is any part of this driven by fashion rather than merit?
3. Solves the right problem
Bad design is often industrious but misguided. The stove dials should match the burner layout, not sit in a row. An intractable problem can usually be replaced by an equivalent one that's easy to solve.
Ask: Is this solving the actual problem, or a convenient adjacent one? Could the problem itself be reframed?
4. Suggestive
A painting that suggests is more engaging than one that tells. Software should give users basic elements they combine as they wish (like Lego). A building should serve as backdrop for whatever life people want to lead in it.
Ask: Does this leave room for the user? Does it dictate, or does it enable?
5. Often slightly funny
Humor is related to strength. The confident often seem to be making fun of the whole process slightly. Hard to imagine something humorless also being good design.
Ask: Does this take itself too seriously? Is there any lightness or wit?
6. Hard
Hard problems force you to toss everything unnecessary. Fashions and flourishes get knocked aside by the difficult business of solving the problem at all. Wild animals are beautiful because they have hard lives. Good pain (running), not bad pain (stepping on a nail).
Ask: Is the difficulty here productive (forcing elegance) or accidental (fickle requirements, bad tools)?
7. Looks easy
Great designers make it look easy — like a Leonardo head drawn in a few lines. You have to get them in exactly the right place. The easy, conversational tone of good writing comes only on the eighth rewrite.
Ask: Does this feel effortless to use/read, even if it was hard to make? Or does the difficulty leak through?
8. Uses symmetry
Repetition and recursion. The Eiffel Tower is a recursive solution — a tower on a tower. Inductive proofs are wonderfully short. In software, a problem solvable by recursion is nearly always best solved that way. But symmetry as a substitute for thought is dangerous.
Ask: Are there natural symmetries being ignored? Or forced symmetries masking laziness?
9. Resembles nature
Nature has had a long time to work on the problem. It's a good sign when your answer resembles nature's. Working from life gives your mind something to chew on. Genetic algorithms let us imitate nature's method as well as its results.
Ask: Does nature solve an analogous problem? What would the organic version look like?
10. Is redesign
Experts expect to throw away early work. They plan for plans to change. Leonardo had five or six attempts to get a line right. The Porsche 911's distinctive back only appeared in the redesign of an awkward prototype. Cultivate dissatisfaction. Make mistakes easy to acknowledge and fix.
Ask: Has this been through enough iterations? Is there reluctance to throw away work that isn't right?
11. Can copy
The greatest masters achieve selflessness — they just want to get the right answer. If part of the right answer has been discovered by someone else, that's no reason not to use it. Unknowing imitation (copying an imitator) is bad. Conscious, confident copying is fine.
Ask: Is there a known good solution being ignored out of desire for originality? Is this imitating the source or an imitator?
12. Often strange
The best work has an uncanny quality: Euler's Formula, the SR-71, Lisp. You can't cultivate strangeness directly — only avoid squashing it. The only style worth having is the one you can't help. There is no shortcut; go through good and come out the other side.
Ask: Is any natural strangeness being suppressed to look "normal"? Strangeness that arises from getting it right should be kept.
13. Happens in chunks
Great work comes disproportionately from hotspots: Florence 1450, the Bauhaus, Xerox PARC. A community of talented people working on related problems is more powerful than individual genius. Being too far removed from a center makes good work nearly impossible.
Ask: Is this work connected to a thriving community or tradition? What are the best practitioners doing?
14. Is often daring
At every period of history, people believe things that are ridiculous. Today's experimental error is tomorrow's new theory. Pay attention where conventional wisdom and truth don't quite meet.
Ask: Is any part of the conventional approach here actually wrong? What would you do if you weren't afraid of looking foolish?
How to Apply
Taste review workflow
- Read/examine the work in full before judging
- Listen for ugliness first — it's easier to see ugliness than imagine beauty. What makes you flinch?
- Run each principle as a lens. Not all 14 apply to every work — focus on the ones that reveal something
- Prioritize — which principles, if addressed, would most transform the quality?
- Suggest specific changes — taste without the ability to gratify it is incomplete
Output format
For each issue found, state:
- Principle violated (e.g., "Not simple enough")
- What's wrong — the specific ugliness
- What would be better — a concrete alternative
End with: what is already good (taste means recognizing beauty too, not just ugliness).
