The Ruins We Already Have


2025, digital animation, live code interface, digital prints, perfume
Collaboration with Hans Tursack
with sound composition by Nick Kopp


Exhibited at Realms Unreal, The Capital Region Arts Center, Troy, NY



In a world increasingly escaping into the imagined futures of science fiction and dystopia, this piece focuses on the ruins that are already here—the fragments of infrastructure and material debris that shape our present. These remnants of abandoned industrial sites, towers, and quarries resist deletion or erasure, demanding attention as sites of both material weight and poetic possibility. Tursack’s animation on the first screen invites viewers to experience the scale and physics of infrastructure—fragments in motion allude to the demolition of obsolete buildings, mechanical systems, and storage facilities. Through these digital physics, a process of material empathy is engaged, where our seeing body reacts to the sensation of falling objects. 

We then used a neural-network-trained program called YOLO ("You Only Look Once") to analyze Tursack’s animation of fragmented landscapes, producing information in the archetypes of AI logic: pattern recognition and probability of "what this is" or "what comes next." This information is representative of how the AI “sees” the animation, how it understands visuals and form. Instead of using this data literally to tell us about the animation, we instead turned its way of seeing towards poetics, using it as a selector for a library of my photographs, archival images and text, which are generated in real time in the exhibition.

The images are mostly photographs reflecting my interest in postindustrial detritus and material of Anthropocene landscapes, but they are also literally tethered to the deeply material infrastructure that powers it, including the reanimation of ruins like Pennsylvania’s Three Mile Island—a nuclear facility (not far from where both artists grew up) slated to re-open in 2028 to power Microsoft’s AI projects. It is also the site of the country’s worst nuclear meltdown in history.







AI is often imagined as an immaterial entity, and both the speed of language models like ChatGPT and usage of terms like “the cloud” provide a veneer of weightlessness. But the cloud is not ethereal; AI has an insatiable demand for energy, increasing our collective ecological risk in a time of desperate need to reduce fossil fuel usage. Microsoft seeks to repurpose the latent energy of nuclear ruins for contemporary digital processes, and in researching its plans, I discovered the ororaburus hiding in plain sight: AI could, in a way, be credited for its own reamination: Microsoft had an AI sort through hundreds of pages of enormously complex nuclear regulatory paperwork, training it to generate the documents needed for a lengthy licensing process to re-open Three Mile Island. Microsoft’s AI system was responsible for securing its own power source—one that could not likely have been done by human lawyers alone.





A hanging sculpture alludes to the material infrastructure upon which AI relies, specifically Three Mile Island nuclear plant in Pennsylvania (home to both artists), which will be re-opened in 2028 by Microsoft to power its AI initiatives. The sculpture was designed and 3D-printed by Tursack in response to industrial design aesthetics and containers for hazardous materials. Inside, a kind of “late industrial perfume” created by Wist combines scents associated with infrastructural ruins (wet concrete, rusty metal), the scent of radiation interacting with the air (reportedly a faint ozone-like smell), and other smells from factories or industrial sites (metallic, burnt plastic). The latter were partially informed by interviews with her parents, both of whom worked in steel mills in Pennsylvania. The sculpture and perfume is meant to remind us of what ruins of the past become re-animated for technology of the future.






Instead of employing AI for utility, we approach it as an active participant in the creation of new aesthetic and conceptual frameworks, and as a collaborator in a poetic exploration. Neural networks, trained for object identification, analyze the infrastructural fragments in Tursacks’ animations. However, their outputs are reoriented—not toward functional categorization, but toward generating associative, fragmentary texts and images that mirror the disjointed nature of these ruins. The AI’s pattern recognition becomes a lens through which to see the fragments anew.

Hans Tursack is a designer from Philadelphia, Pennsylvania. He received a BFA in studio art from the Cooper Union School of Art, and an M.Arch from the Princeton University School of Architecture where he was the recipient of the Underwood Thesis Prize. He has worked in the offices of LEVENBETTS Architects, SAA/Stan Allen Architecture, and Tod Williams Billie Tsien Architects. His writing and scholarly work have appeared in Perspecta Journal, Pidgin Magazine, Plat, Crop Journal, Thresholds Journal, Log Journal and Acadia. He was the 2018-2021 Pietro Belluschi Fellow at the MIT School of Architecture + Planning and a Visiting Assistant Professor at the Sam Fox School, Washington University in St. Louis (2022-2023). Hans is currently a PhD student in Electronic Arts at the Rensselaer Polytechnic Institute’s Department of Humanities, Arts and Social Sciences.