Self-Generating City: Training an AI model on a corpus of glitched 3D prints
Image of a city: generated with a model trained from scratchThey are transitional objects and this need for transition often turns into madness. So one might therefore wonder, for every city, what psychological state led to the accumulation of so many things, buildings, crossings, cars, relationships and people. What neurosis presided over the births of different cities?

This project explores the fetish object of a “failed” 3D print. Created out of disregarded 3D prints from the 3D print lab, the self-generating city in essence builds itself.

Detail Imagery Used to Train
What can glitches tell us about the future of architectural adaptation?

Tower for a Collector, Studio MUOTO
Professor: Gilles Delalex, Yves Moreau