AI Impacts Design & Architecture | Molly Wright Steenson | Talks at Google

Molly Wright Steenson is the K&L Gates Associate Professor of Ethics & Computational Technologies at Carnegie Mellon University & Senior Associate Dean for Research in the College of Fine Arts. A designer, writer, and international speaker, she is the author of Architectural Intelligence: How Designers and Architects Created the Digital Landscape (MIT Press, 2017), which traces the radical history of AI’s impact on design and architecture, and the co-editor of the forthcoming book Bauhaus Futures (MIT Press, 2019). Molly cut her teeth on the Web in 1994 and has worked at groundbreaking design studios, consultancies, and Fortune 500 companies. She holds a PhD in Architecture from Princeton University and a master’s in architectural history from the Yale School of Architecture.

In this talk, Molly Wright Steenson explores the work of four architects in the 1960s and 1970s who incorporated elements of interactivity into their work. Christopher Alexander, Richard Saul Wurman, Cedric Price, and Nicholas Negroponte and the MIT Architecture Machine Group all incorporated technologies—including cybernetics and artificial intelligence—into their work and influenced digital design practices from the late 1980s to the present day.

[vp]

Christopher Alexander is not enthusiastically embraced by architects and planners is probably due to his overselling of the pattern language. His aims were much larger than combinatorial flexibility and modularity. He promised a new urban theory. His 253 patterns were inadequate in the face of modern urbanization with mixed uses and complex built forms. In practice, the pattern language fell short. In theory, the pattern language was inadequate. Hence the disappointment. In fact this disappointment became apparent and Alexander himself noted in the very first project he undertook for one of the neighborhood plans in San Francisco.

A similar disappointment came about in computer languages that mimicked patterns in object oriented languages (I'm mainly thinking about gang of four). Patterns fell woefully short when expanding methods and classes. Large-scale software projects became unmanageable and a crisis ensued. Functional languages, such as Haskell, solved some of these problems in earnest for the first time, and a major crisis averted. Now older imperative languages such as C and Java are retrofitting functional programming concepts such as lambda calculus, lazy evaluation, monads, and so on. Patterns in software, like in the brick and mortar world, are no longer seen as the panacea for modularity.

After decades of false starts, we have realized that modularity is not about finding connections in the superficial methods or properties or features. It goes much deeper to the very the functional aspects. In software, it is found in deeply embedded mathematical operations. In architecture, it is found in deeply embedded aesthetics.