Future Follies: Design Logic in the Age of GenAI-Modelling and Real-Time Visualization
Vision-language AI models can produce precise 3D geometry when coupled with CAD software. Novel rendering techniques such as Gaussian splatting enable the seamless embedding of a design within an immersive environment. Participants will explore the architectural potential that emerges when these two technologies are combined. Can AI-generated geometry adhere to explicitly defined design constraints? What contextual insights can be derived from a rich visualization? The task is to extend Bernard Tschumi’s seminal work on architectural follies by generating new designs for a site at ETH that follow Tschumi's rule-based logic. State-of-the-art tools will be provided through cloud-based virtual machines.
Tutors
Lidor Gilad (Itten+Brechbühl AG, Zurich / Accademia di Architettura, USI, Mendrisio, Switzerland)
Lidor Gilad is Head of Design Switzerland and Partner at Itten+Brechbühl. He is a senior lecturer and researcher at the Accademia di Architettura / USI Mendrisio, directing the course on Digital Representation, which teaches 3D-modelling and rendering software as a tool for morphological, structural and conceptual analysis and synthesis of architecture.
Pierre Chevremont (Itten+Brechbühl AG, Zurich, Switzerland)
Pierre Chevremont is Head of AI at Itten+Brechbühl. He graduated from ETH Zurich in Architecture and holds an MSc in Cognitive Science. He has led several AI workshops at USI Mendrisio, TU München and ETH and co-founded the art project Constructed Visions, which explores the potential of AI to visualize possible architectural worlds.
Patrick Bedarf (Itten+Brechbühl AG / ETH Zurich, Zurich, Switzerland)
Dr. Patrick Bedarf is AI and Digital Technology Expert at Itten+Brechbühl and a senior lecturer at ETH Zurich. He specializes in digital design and fabrication technology research that connects academia with practice and has led numerous international workshops.
Nora Walter (Nora Walter Images, Zurich, Switzerland)
Nora Walter graduated from TU München in architecture, specialized in visualization at Caruso St John Architects and founded Nora Walter Images in 2020. Since 2021, she has regularly led workshops on the digital representation of space, light and materiality, adding generative AI in 2024. She co-founded Constructed Visions in 2023.
Elia Rossi (Accademia di Architettura, USI, Mendrisio / Superfluido Productions, Zurich, Switzerland)
Elia Rossi develops next-generation architectural workflows that integrate interactive design, real-time visualization, and AI-enhanced immersive VR. Based in Zürich, he brings together technological innovation and academic research, leading Superfluido Productions and V-ARC while lecturing on Integrated Digital Tools, Digital Fabrication, and BIM at USI Mendrisio.