Diffusions in Architecture: Artificial Intelligence and Image Generators
by Matias del Campo, Lev Manovich
Epilogue
Matias del Campo
There are probably a million ways to end this book, but let me end with a little story. Walter Gropius could not sketch. At all. Not a bit! At the start of his career working in Peter Behrens’ office, Gropius kept this handicap a secret.1 Despite this setback, Gropius persevered in the Behrens office and eventually became known for his ability to dictate drawings to his collaborators, proving the power of language to convey the intricate material and spatial relationships in architectural projects.2 This notion of description was further expanded upon by Sol LeWitt, who dedicated his entire career to the potency of language in the form of instructions, presaging the emergence of programming and scripting as a means of artistic and architectural expression.3 The synergy between language and architecture is complex and multifaceted, yet it is clear that language has the ability to convey abstract concepts and facilitate collaboration, offering a powerful tool in the design process. To this end, this book attempted to shed light on the murky shadows of diffusion models by exploring their relationship with language and the architectural imagination. As we close the last chapter of this book on diffusion models in architectural design, it seems unmistakable to contemplate on the ways in which language, prompting, and the theoretical perceptions of thinkers like Wittgenstein, Roland Barthes, and Michel Foucault have fashioned our views upon technologies ...