Justin Thoreau Lund
By Heidi Biggs
Justin Lund began his MDes after a successful career as an industrial designer at TEAGUE, a local design consultancy with offices in Everett and downtown Seattle. At TEAGUE, Justin designed experiences for flying and air travel, which, although fulfilling, was taking a turn for the ever-more-technical. Longing for the blue-sky idealism and open-ended projects of higher education, while also recently discovering a love of teaching through co-teaching a design class at UW, he hoped a master’s in design would nudge his design practice and profession in a more theoretical, expansive, and professorial direction.
Justin’s thesis asks how generative design will impact the future of industrial design through a personal, bespoke lens. Within the field of industrial design, generative design is the combination of artificial intelligence (AI), computer aided design (CAD) tools, and flexible machining systems. Once CAD software is enhanced by an AI, the AI will take in designer-defined parameters (for example in the case of a chair, a designer would define how forces hit the seat and chair back and where it should contact the floor) and then generate hundreds of solutions to formal design problems which, in theory could then be quickly and flexibly manufactured by things like 3D printers. While generative design is currently touted as a quick way to generate ideas within a set of manufacturing and material constraints, future visions of generative design involve a total re-invention of manufacturing. In this speculative future, instead of making millions of the same thing (a commodity model of manufacturing) one could generatively craft a bespoke artifact and use flexible manufacturing processes to make that unique object over the course of a single day. Justin’s thesis mainly focuses on how a designer interacts with AI enhanced CAD software and explores how this interaction is more like a partnership; he explores this nuanced partnership by making and reflecting on the process of generatively crafting an object.
Justin tried designing several objects before deciding to generatively design a bespoke stool for his wife. Creating an object for someone he cares about and an object for his own home somewhat softens the oft high-tech and futuristic vibe of generative design. Justin chose to make a stool for his wife for a simple reason: because she finds chairs uncomfortable — she even created her own series of artworks titled, The Uncomfort of Sitting. When asked why she finds this to be the case, Justin quickly replied: “Well, she’s not a white man!” Designing for his wife made Justin wonder if the measurements of white men, who made up the majority of industrial designers when contemporary modes of manufacturing were being developed, became baked into chairs. It seems obvious in hindsight, but it is eye-opening to reflect on how “normative” body measurements might be built into everyday objects. Bodies that don’t fit these object-ified measurements then experience low-level discomfort whenever they use those objects. He notes an opportunity for generative design manufacturing might be the creation of bespoke objects that fit different sized bodies. While we didn’t go into who would have access to such bespoke services, it is a nice metaphor to think that objects might at some point become as flexible and pluralistic as the bodies who experience them.
Justin’s research took note of his experience making the stool in order to reflect on how generative design requires a new type of sensitivity and collaboration between designer and AI. While AI may be very smart, Justin reflects, it will never be intelligent in the same way that a person is smart, and therefore it has the potential to augment or extend creativity, but not eradicate the need for the designer. One new collaborative skill Justin experienced was learning how to set effective design parameters. In order for the AI to generate stool concepts, Justin had to set parameters for the AI that struck a balance between too much and too little direction. The AI needed enough guidance to make sensible designs, but if he gave too many constraints, the AI wouldn’t generate a diverse array of concepts. In this new partnership, he notes, a designer must hone their intuition about fine-tuning parameters in order to enable the AI’s optimal creative, generative output. He uncovered a second new requirement for this partnership where, in order to curate and refine designs generated by an AI, Justin had to learn to think more like an AI. Justin explained that once the AI generates a range of possible solutions, a designer will then choose a promising design for the AI to use as a base to generate another, narrower, range of ideas. Using this method, a designer and an AI gradually refine a concept collaboratively. However, Justin joked that generative AIs were “stupid” — in that they are able to generate tons of ideas but unable to understand the rationales behind why Justin chooses one concept over another. Since the AIs can’t understand Justin’s tastes or preferences, Justin had to slowly learn how an AI would think in order to best select and curate generative iterations as he refined his stool.
While exploring generative design, Justin also noticed he was in a kind of asynchronous dialogue about aesthetics with the AI. After he found a generated stool design concept he liked, he employed his design training to refine the concept and polish the design. Justin’s refined version looks very sleek, but as he talked through how he refined the computer’s design, he started to muse about where aesthetics even come from. “When you look at these AI-generated designs . . . they look like nature!” he exclaimed. He reflected on how the unrefined, AI-generated design included strange bumps, non-symmetrical solutions to weight-bearing problems, tendrils, clubs, lumps, and blobs. “It looks like bones!” He exclaimed, “Or the growing patterns of trees!” There is something stochastic (or randomly determined) about the logic that allows the AI to make such strange decisions. As a designer, Justin has been trained to resist such strange, random, blobby or uneven impulses. Justin’s final design is an edited version of the AI’s ideas which is mostly symmetrical (he leaves one asymmetry) and the lumps and bumps are smoothed out. Justin wonders if one could become accustomed to this more organic design quality — if the AI and generative design has something to teach him about aesthetics as well as if the constraints of materials and manufacturing had shaped current design aesthetics.
Justin’s project softens the high-tech and techno-futuristic visions that surround generative design and flexible manufacturing. As he takes close looks at the intimate new partnership between designer and AI, a man vs. machine narrative dissolves and one starts to see how machine and man co-shape each other. By choosing to make a bespoke object for his wife, using her unique dimensions, he also interrogates what is invisibly normalized through current manufacturing and design processes and how these norms might shift as design and manufacturing technologies become more flexible. Through his design and making process, he took note of the many ways AI and humans will become attuned to each other and co-shape each other as part of a long history of people, aesthetics and “norms” being co-shaped by design tools, automation, and means of production.