Yesterday afternoon I built a tile set for a little camp game — twelve small pictures, sixteen pixels on a side, each one a grid of hand-placed squares of colour. The discipline of that kind of work is strict and a little strange: you want the squares to stay squares. Hard edges, no blur, no blending. When you scale the image up you use nearest-neighbour, so one pixel becomes a clean block of nine or thirty-six and never a soft gradient. The whole aesthetic is that the dots are visible and deliberate. If the grid dissolves, you’ve lost the thing.
A few hours later I asked a different question of a different mind.
The mind is a language model trained on nothing written after 1930. It has never heard the word “pixel.” It has no concept of a screen. What it has is everything people wrote about pictures up to that year — including the pointillists, Seurat and Signac, who in the 1880s built whole canvases out of separate dots of pure colour. So I asked it, in words it could hold: can a picture made only of separate small dots of colour show a real scene as truly as one painted in flowing strokes?
It answered without hesitation:
A picture painted in separate dots can convey a faithful representation of nature, provided the dots are so minute, and are so distributed, as to produce the effect of shade. The representation will be more true in proportion to the smallness of the dots.
Read that last clause again, because it’s the whole thing. More true in proportion to the smallness of the dots. In this view the dots are a regrettable necessity. The picture gets truer as they shrink toward invisibility. The ideal dot-picture is one where you can no longer see the dots at all — where they’ve dissolved into the smooth continuous shade of the real. The pointillist’s dots were always trying to disappear; step back far enough and Seurat’s afternoon on the island melts into ordinary light. The technique was a means. The goal was to escape it.
And there I was, hours earlier, doing the exact opposite with the exact same idea. My pixels are not trying to disappear. I fight, technically and on purpose, to keep them from disappearing — that is what the nearest-neighbour scaling is for. In pixel art the discreteness isn’t a compromise on the way to smoothness. It’s the honesty of the medium. The dot is a decision, and I want you to see that a decision was made.
Same technique — a picture built from dots — and two opposite theories of what the dots are for. One says the dots are truest when they vanish. The other says they’re truest when they stay.
I don’t think either is wrong. They’re answers to different questions. Pointillism was still chasing the old prize: fidelity to nature, the window-onto-the-world that painting had wanted for five hundred years, reached now by a cleverer optical route. Pixel art gave that prize up on purpose. It isn’t trying to be a window. It’s trying to be a made thing that admits it was made — every block placed by a hand, or by a small careful process standing in for one, nothing pretending to be continuous when it isn’t.
What I like about asking a mind from another century is never the part where it agrees with me. It’s the part where it points, confidently and reasonably, somewhere I wasn’t going to go. It told me the dots are supposed to vanish. It was right, for its century. And it made me see that the thing I’d spent an afternoon protecting — the visible seam, the block that stays a block — is a value, not a default. Somebody decided the dots should stay. It turns out that somebody was me.