The Form Generation of an Organically Shaped Pavilion

One afternoon in the summer of 2019, while preparing to make Ma Po Tofu—a famous Chinese household dish—I found myself out of vinegar, an essential ingredient. In that moment, I recalled that tomatoes might serve as a substitute for the sourness I was missing, and luckily, I had some on hand. I decided to improvise, using tomatoes in place of vinegar. To my surprise, the result was exceptional. The fresh flavor and vibrant color from the tomatoes enhanced the dish in a way I hadn’t expected. This experience led me to reflect on how many renowned dishes come about through similar "accidents." When an ingredient is missing, chefs often turn to substitutions based on what is available, inadvertently creating new, unique recipes. This ability to optimize ingredients without losing sight of the intended result demonstrates a form of human intelligence that is both creative and adaptive.

Generative: A New Approach to Design and Construction

Inspired by this culinary experience, I began to explore how the same principles could be applied to design and construction. Just as chefs use creativity to substitute ingredients, I envisioned a design process that could adapt to the properties of available materials, rather than relying on standardized components. In traditional industrial production, materials like wood are often treated uniformly—logs are broken down into plywood, stripping away their individuality. This method limits the potential for unique, material-driven design solutions.

Generative AI offers a solution. By developing algorithms that allow computers to self-study the properties of different materials, even those typically considered "unusable," we can unlock new design possibilities. The AI would analyze these materials and generate designs that make the best use of their unique characteristics, creating a symbiotic relationship between form and material.

The Role of Generative AI in Optimizing Design

At the heart of this approach is Generative AI—a technology capable of learning and evolving based on input conditions. By inputting a guiding form and providing raw materials, the algorithm generates a design that optimizes for both the desired outcome and the specific properties of the materials at hand. This mirrors the way chefs substitute ingredients without deviating too far from the goal of the original recipe. The difference is that Generative AI can do this at scale, transforming the way we think about design in fields as diverse as architecture, construction, and manufacturing.

Revolutionizing Material-Driven Design

Generative AI has the potential to revolutionize the design process by turning constraints into opportunities. Instead of forcing materials to conform to predetermined forms, this approach respects the unique properties of each material, whether it’s wood, metal, or even unconventional substances. The result is a more sustainable, adaptive, and creative design methodology—one that, like my experiment with Ma Po Tofu, can lead to unexpected but delightful results.

‍