Aug 14, 2026
There is a particular pleasure in standing at the edge of a new medium before it has quite decided what it is.
I have spent much of my working life moving between disciplines that were once treated as separate territories: graphic design, photography, illustration, film, sound, animation and interactive media. Each has its own grammar, its own machinery and its own way of persuading an audience to believe in something that isn't there. A photograph fixes a moment. Cinema stretches it through time. Graphic design compresses an idea into composition, colour and form. Sound can make an empty room feel inhabited.
Generative AI seems to be dissolving the borders between them.
I approached this un-momento with that thought in mind, using open-weights MinMax H3 as a kind of virtual test camera. I wasn't particularly interested in asking whether it could produce a convincing science-fiction image. That question has already become rather dull compared to the craft itself, what people produce. The more compelling question to me is whether it could begin to participate in the process of directing an image: establishing a world, controlling its visual language, placing a character within it, and imagining how a camera might travel through the scene.
Is this toolset capable of undestanding my original concepts?
A giant landing platform stands watch above a city at sunset, so large that its underside becomes almost architectural. Its industrial hull occupies much of the sky, its surfaces disappearing into shadow and catching the last orange light of the day. Below it, a lone figure stands at the edge of a metal platform, looking out across an immense city buried in atmospheric haze. Thousands of lights shimmer beneath the setting sun. Far away, a needle-like tower rises from the haze.
What immediately struck me was not the complexity of the machinery, although there is plenty of it. It was the sense of scale. Atmospheric perspective. Accurately-appearing scale. Structures in the distance catching the sunlight. A colour palette that works to evoke emotion. Compositional weights deftly placed throughout the camera move. Compositional narrative that makes sense.
The image has three worlds in it. There is the human scale of the man, the technological scale of the spacecraft, and the almost incomprehensible scale of the city below. The figure is reduced almost to a silhouette, yet he becomes the emotional centre of the photograph precisely because everything around him is so enormous.
That is something I have always looked for in visual storytelling: the point at which scale becomes meaning.
The machine isn't simply a machine. The city isn't simply a background. They tell us something about the person standing between them.
I began to wonder what would happen if the still image were treated not as an illustration, but as the opening frame of a film.
Generative systems have historically been very good at producing moments and much less reliable at maintaining continuity between those moments. They can give you a remarkable character, and then give you another remarkable character who happens to be wearing the same clothes. They can construct an extraordinary spaceship, only to redesign it when asked to look at it from another angle.
For a graphic designer or filmmaker, this is a fundamental problem. We don't simply need images; we need relationships between images.
The character must remain the same person. The architecture must remain the same architecture. The light has to belong to the same sun. The spacecraft has to occupy the same physical world.
In these experiments, I found myself pushing MinMax less like a prompt box and more like a visual collaborator: describing not only what should be present, but the way the scene should be seen. The distinction sounds small, but it changes the creative process. I am no longer merely describing an object. I am beginning to describe a shot.
I have always been fascinated by light because it is simultaneously physical and emotional. In photography it tells us where we are. In cinema it tells us how to feel about where we are. In illustration it can become almost symbolic.
Here, the colour contrast creates two opposing worlds without requiring anything to be explained.
The technology is cold.
The horizon is warm.