At its simplest form, a play button is really just a triangle inside of a circle. But if you center that triangle mathematically, dead center of the circle, it looks wrong. It looks like the triangle has slid to the left. So, you nudge it a few pixels to the right (which is off the true center), and now it LOOKS centered. Every designer has done this… most of us stopped noticing we do it.
That small correction is closer to the real job than people think. Design was never really about arranging shapes according to math. What you’re actually designing is the perception of the geometry (not the geometry itself), and human perception is full of quirks that pure geometry knows nothing about.
Take a capital O and a capital H set to the same height. The O will look smaller, because your eye shrinks curves. So type designers draw the O taller, pushing it slightly past the cap line and below the baseline, until the two look equal. The letters that appear to match are the ones that don’t measure the same.
A circle and a square sitting in the same icon grid have the identical problem. If you match their bounding boxes, the circle looks undersized. So you scale it up until it stops looking wrong.
The optical center of a page sits higher than the mathematical center. If you put a title at the exact vertical middle, it looks like it’s sagging toward the floor. So you move it up a touch, and it settles.
Spacing works this way too. You can’t space letters by putting equal distance between them, because the gaps are different shapes and your eye counts area, not distance. An A next to a V wants to tuck in close. Two straight-sided letters want room. Kerning is really just the practice of making unequal distances look equal, done entirely by eye, one pair at a time.
Futura is my favorite example, because the whole typeface is basically a confession. It’s sold as geometry, the typeface of pure circles and triangles, and it reads that way. But put a ruler on it and almost nothing survives. The O isn’t a true circle, the strokes thin where a curve meets a stem, the pointed apex of the A overshoots the cap line. Renner drew the corrections in because the honest geometry looked wrong.
Then there are logos. People love the diagrams that lay a famous mark over a perfect scaffold of circles and golden rectangles, as if the thing were derived from arithmetic. Most of those diagrams are drawn after the fact. The real mark almost always breaks the grid somewhere, because the version that sat perfectly on the circles looked slightly off, and someone with a good eye pulled a curve out of alignment until it looked right. I’ve made that pull on more marks than I can count. Nobody has ever noticed… which is the point. You’re supposed to see something that simply looks correct, without ever knowing that someone had to break the grid to get it there.
Now notice what’s actually happening in all of these. The shapes were never wrong. Our eyes are, and they’re wrong in specific, predictable ways. The whole discipline is built around that flaw.
Which is exactly why AI-generated design so often feels a little ‘off’.
A model doesn’t see its own output. It has no eye that shrinks a circle, no discomfort when a title sags. Trained on enough human-corrected work, it’ll often draw the O taller than the H, because that’s how the type it learned from was built. It’s copying the outcome without the instinct. It never felt the illusion, so it has no reason to fix one it hasn’t already seen.
That gap shows up two ways. Sometimes the work is mathematically clean and reads as wrong. An icon centered to the pixel that still looks shoved left. Two shapes sized to the same box where one looks shrunken. Everything measures correct and nothing looks correct, because measuring correct was never the goal (it was just the part you could write down).
The other way is quieter and more common.
A model averages toward the middle of what it has seen, which produces something competent and faintly lifeless. It won’t make the specific, intentional break a person makes on purpose, because it can’t have the reason. The reason was a human looking at the screen, feeling that something was a hair too tight, and moving it. The machine has nothing to feel with.
This is the part people miss when they assume better models will close the distance on their own. The known corrections are learnable. They sit in the pixels of every well-drawn typeface, which is how a model comes to copy them. What’s not in the pixels is the reflex that produced them: a specific piece of biological hardware getting fooled and pushing back. A model can inherit every recorded fix and still be helpless in front of a new composition that needs one nobody has drawn yet. Until it has something that works like being fooled, its design will keep needing a person to look at it and say “move that two pixels to the right.”
So there’s a larger claim folded inside this small one. A spec is supposed to describe the thing being built. Dimensions, weights, hex values. But every correction I’ve shown you exists because the real acceptance test was never the measurement. It was a person looking. The deliverable was the perception all along… we wrote our specs in geometry because geometry is what fits in a document.
That decision still belongs to a person. The eye is the spec.