Between (re)drawing and representation across generative models

Drawings, physical models and words have long been the traditional media through which we represent, communicate, translate and record our experience of architecture. The shift into the digital (software models, parametric design, immersive technologies) has not only extended these modes of representation but fundamentally changed our visual and formal vocabulary.

The recent rise of large language models (LLMs), image diffusion models, and multi-modal generative models raises a critical question: what happens when we rely on such models to represent architecture? What happens when we offload our understanding of spatial experience to an algorithm, and then adopt the visual vocabulary of a diffusion model as architectural truth?

This research investigates how machine learning models are transforming architectural representation, and how AI-generated outputs alter our understanding of space, memory and experience. It explores these questions through two ongoing interconnected works by the author, Anamnesis House and Indexical Drift.