1st Edition

Compositional Intelligence Architectural Typology Through Generative AI

By Daniel Koehler Copyright 2026
256 Pages 93 Color Illustrations
by Routledge

256 Pages 93 Color Illustrations
by Routledge

256 Pages 93 Color Illustrations
by Routledge

*With a foreword by Mario Carpo* Given the extensive data sources represented by urban environments, this book investigates the potential of architecting for cities using large-scale models – the technological underpinning of Generative AI. Koehler examines Large Language Models (LLMs) in relation to architectural questions of typology – those critical junctures where architecture engages... Read more

1. Introduction - The City as the Affordance of Language  2.The Digital City and the Erosion of Human-Centric Design  3.Out of the Blue: Computing Beyond the Virtual  4.The Generated City  5.Model: From Representation to Synthesis  6.Language to Latent Space: Architecture’s New Materialism  7.Large: The Architecture of Scale  8.From Dialectics to Superposition:
Architectural Implications of Computational Criticality  9.Architecture as an Embedding of Languages

Biography

Daniel Koehler is an urbanist, architect, and Assistant Professor of Architecture Computation at UT Austin. His research, previously conducted at the Bartlett in London and Innsbruck University, led to his first book The Mereological City (Transcript/Columbia Press), examining part-relationships in modernist urbanism. Koehler’s work has been exhibited internationally and is included in the Centre Pompidou’s permanent collection. He serves on the ACADIA board of directors and the IJAC editorial board. At the intersection of architecture and computation, his research navigates the paradoxes and potentials of AI and digitalization, developing typological interventions that reveal environmental, economic, and societal synergies.

"This book by Daniel Koehler is one of the most comprehensive, ambitious efforts to date to come to terms with Generative AI's baffling capacity to invent new images that are recognizably similar to all the exemplars from which they derive, yet identical to none--and to infer from this epistemological conundrum both analytic tools and a viable design strategy"

Mario Carpo