Cosmic Composites and Prehistoric Perception: How One Artist Brings Light to the Connections Between AI, Space, and Creativity

Sam Price
6 min readApr 14, 2023
“who are we where are we” (Sam Price, 2023), Projection on Polystyrene, 8’ x 8’ x 1.5’, 40:00.

In 2022, a flurry of advances in machine learning and space exploration dominated public attention. With the former, decades of academic research and expert coding yielded the first widely adopted image-generation platforms, which would enable anyone from all backgrounds to generate imagery from text. In the case of the latter, a record number of satellite and rocket launches provided window-dressing for the highly complex, almost improbable success of the James Webb Space Telescope, humanity’s most advanced cosmic lens.

As these landmark technological achievements push the micro-and-macro limits of human visual limits, it is worth reexamining what preceded their inception: myth-making. In one of its earliest stages, this likely took place around the hearth of a roaring flame, lighting up dark surroundings and a night sky pregnant with uncertainty. It was here where humanity’s primal ancestors assigned meaning to the cosmos and their surroundings. They made further sense of their environments by projecting their perceptions onto cave walls.

Photo of the Lascaux Caves’ interior. Credit: Bayes Ahmed. Uploaded by Jan van der Crabben, April 27, 2018.

This animistic imagery later took to the stars in the form of constellations. The earliest suspected depictions exist in the caverns of Lascaux, France, within a labyrinthine collection of infamous 17,000-year-old cave art. Beyond their spiritual significance, such clustered associations were used for more functional purposes like navigation and tracking the passage of time.

A section from “The Caves of Altamira” (Sam Price, 2021) Acrylic, 48" x 31"

Today, humanity’s perceptual appetite is satiated by two novel visual technologies: AI that can conceive imagery from inner space, and increasingly advanced telescopes that can clarify the depths of outer space. Yet, at the center of this inquiry is a timeless, almost-clichéd search for the truth of what we really are. What underlies the organization of inanimate minerals into enlightened consciousness, that, in turn, aim to more directly materialize their perceptions into their environments?

Inspired by these extensions of human perception, multimedia artist Sam Price created a sculptural projection that integrates both. “who are we where are we” is a testament to the instinct to seek patterns among the stars, and humanity’s continued effort to embody its visions.

Underlying Price’s art practice is “eigengrau” (German for intrinsic grey), which signifies the color one sees through closed eyes in the dark. Price was originally drawn to the term during his BFA thesis research:

“What I found so interesting about eigengrau is that it inherently precedes every dream and many great ideas. It is a universal yet almost ineffable experience. As such, I viewed it as the ultimate creative blueprint; it now plays a role in the process, display, and conception behind my artwork.”

“Eigengrau,” (Sam Price, 2020) Spray Paint and Plaster on Foam; Wire, Stepstool, 3’ x 3’ x 4’

Price explains how his interest in AI and space — and their connection to the phenomenon of eigengrau — informed his conceptualization of WAWWAW:

“The eigengrau experience parallels our evolutionary origins, in which our ancestors’ uncharted surroundings may have spurred sentience out of a need for survival. Machine learning algorithms take the eigengrau experience to the next level; they are black boxes that augment our primeval pattern recognition abilities — enhancing details that we would never notice. In another sense, eigengrau also parallels the act of exploring space, of bringing light to the vacuous cosmos.”

A detail shot of one of the sculpture’s sections during a test hang.

WAWWAW is far from the first sculptural projection by Price. This concentration of work is partially inspired by archaeological studies of prehistoric cave etchings, nicknamed “spaghetti lines’’ by researchers. To the untrained eye, these features, often cut underneath paintings of local fauna, seem intentional, but the reasoning for their placement is unclear. It is only when a light source moves through the space that the animals depicted appear to shift, emerge, and disappear.

“In flickering lamp light, the ibex will appear to [shift] from a grazing position (head down) to a vigilant stance (head up).” Illustration by Nancy Ventura. Learn more about Edward Wachtel’s research into cave etchings.

Price’s project is his second collaboration with Ponce Lab, a neuroscience laboratory located at Harvard University. He primarily worked with Ph.D. candidate Binxu Wang in coding a neural network to seek out lifelike patterns in images of space — mirroring humanity’s eternal quest to characterize the imagery of the night sky.

Wang describes the technical process behind the program:

“In order to achieve our goal, we leveraged a CycleGAN, an AI program that learns to translate between two different visual styles. In our case, we trained the program on two sets of visual data: photographs of life and of space. The goal of our project was to map the style of the space imagery onto the life imagery. The learning process is similar to a game of telephone: images of one style are translated to the other, and then translated back again, ensuring that no content is ‘lost in translation’. An AI judge in the program evaluates each translation to determine if it can be deceived into believing that it is a genuine representation of the target style. This reinforces a recursive process leading to continuous improvement in the translation quality. After carefully adjusting the spatial scale of the image mapping, we obtained some amazing translations.”

A sampling from hundreds of composite outputs generated by the CycleGAN. Watch a preview trailer here.

“As we exported hundreds of composite images, it felt like we were visualizing modern constellations in the night sky — rendering the mind’s eye of ancient cosmic cartographers,” Price added.

“who are we where are we” (light mode), Polystyrene, 8’ x 8’ x 1.5’.

The final images were packaged into a forty-minute projection. Accompanying the video is an eight-foot-diameter disc-shaped sculpture, created from a heightmap of the first James Webb Space Telescope image. Hanging in the darkness, viewers are immersed in a cosmic array that is at once primeval, uncanny, and timeless.

On July 12th 2022, the first JWST image provided a promising preview of the visual power of the new satellite.

WAWWAW will be one of 10 works displayed in Price’s first solo show of 2023, taking place in July at Modfellows Art Gallery in Nashville, TN.

Learn more about the project on, and keep up with Price’s latest work on

Sam is a Boston-based artist who graduated from Cornell with a BFA and Architecture minor. His art has been featured in Cornell’s Herbert F. Johnson Museum, Perime Art Gallery, Miami Art Week, and one of 1stDibs’ first cryptoart exhibitions. Additionally, he was accepted into the 2020 Best of SUNY and SUNY Chancellor’s Gallery Exhibitions. Sam was selected for Cornell’s 2020 Anderson Ranch Painting Scholarship, a \art grant, BitBasel’s CryptoArt for Impact and Innovation Challenge, and the 2020 Edith Adams & Walter King Stone Award recognizing promising work in advance of his thesis year. He had the honor of being selected for the first digital art collection on the moon, contained in a nickel disc aboard NASA’s first return mission in 50 years this June.



Sam Price

Sam is a Cornell 2021 BFA and multimedia artist who explores pareidolia, surreal automatism, dreams, and artificial neural networks in his work.