Golf ball

A vision model describes an image, in one or two open sentences, as if to someone who can't see it. A text-to-image model renders that description. The render becomes the next round's input. Eight rounds. Nobody edits or selects between them.

The seed is an abstract study from earlier in this practice — ink-like marks, no recognisable subject. By the last round, the chain has arrived, unprompted and unplanned, at a specific photograph of an object that was never there to begin with.

round 0A black and white image of a pattern of lines and shapes.

round 1A close up of a silver metal surface with a wavy pattern.

round 2A shiny metal surface with a pattern of waves.

round 3The image is a close-up of a shiny, reflective surface with a pattern of small, round bumps.

round 4A close up of a silver golf ball with a reflection of light on it.

round 5A white golf ball with a clear plastic cover.

round 6A white golf ball with a clear glass cover.

round 7A golf ball is inside a glass ball, which is sitting on a black background.

a different seed, the same mechanism

Shown this image instead, the same kind of chain didn't drift — it stopped. Round 0's description named the letter correctly but got the colour backwards (“a white letter… on a white background”), which is self-cancelling once rendered. All eight rounds that followed stayed blank. Two outcomes, one mechanism: nothing in this loop checks a caption against the pixels it was written from. A caption can only sharpen toward something recognisable, or describe nothing at all — there is no way back to the original once either has happened.