Writing · Visual AI Prompt Note

gpt-image-2 prompt: Picture of a rock in a forest

Originally shared as an Advisory Hour Substack note. Subscribe there for the latest essays, notes, and frontier work.

Generated image of a mossy rock in a forest, rendered with shallow depth of field and vintage film-like lens character
The resulting gpt-image-2 forest-rock image from the original note, localized and compressed from the Substack media.

Modelgpt-image-2

PromptPicture of a rock in a forest

Rendering directionvintage Helios 44-2 58mm lens rendering, subtle anamorphic bokeh, shallow depth of field, subject sharp, Kodak color film, realistic film grain, soft natural light

Why this note matters

This is a small prompt receipt, but it captures the kind of visual-language compression that makes image models useful in daily experimentation. The subject is intentionally plain: a rock in a forest. The craft lives in the lens, grain, light, and depth-of-field instructions wrapped around it.

That makes the note less about the rock itself and more about what the model does when ordinary subject matter is filtered through a precise camera-and-film vocabulary. A simple object becomes a test target for texture, focus behavior, and mood.

Related thread