One potential use case for AI Image Generation | part 1
How much time do you really need to spend on building the story of a project?
The rendering and representation consume a significant number of hours in the early stages. The design changes alone are a big part of that.
This is an exercise in templating/linking and automation and some form of brute force mining for prompts and parameters
It starts with a simple low poly model, that expresses some design intent, it covers at least:
Horizontality, central axis, verticality, a hierarchy of shapes, an order in scale and proportions, but no detail
As close as it can be to a minimum viable 3d model
Setting up the parameters not super straight forward but getting faster
Ran 400 iterations
Picked 64 of the promising ones
They can inform further design intent, act as a place of inspiration for 3d modeling
Higher res bellow
Some final thoughts:
It doesn’t need to be perfect, it needs to be diverse
It doesn’t need to be functional, it needs to map out preferences and advance further exploration
It needs to help in identifying areas that are not important
If we could consider the diagram bellow as a possible representation of the design process (it is still an oversimplification but trying to get the gist of it) then:
The faster we can iterate at every stage ( per vertical column) the faster you can move from one stage to another (one column to another)
