I’ve been playing around with some automated terrain and city generation this morning – a (very) quick experiment I thought I’d share. Smart maps that build themselves based on a set of rules – inspired by something similar from Gemini themselves a week or so ago. For this run, I used Sydney’s geography as the starting point as well as some clean slate options. The system automatically fills in the landscape – placing mountains, forests, and water where they make sense – and then uses Gemini AI to turn that abstract grid into a realistic satellite image in real-time.
Taking me back to my PhD in participatory analytics. Imagine using these kinds of rapid, generative tools for co-design – allowing communities to place in their priorities, generate and sketch out ideas, vote on layouts, and instantly see them rendered as realistic environments.
If anyone else is experimenting with procedural mapping or generative AI in this space, I’d love to hear from you.
Category: Uncategorized
-
Procedural terrain to AI-generated satellite imagery