The Side Valley marks the artist’s first systematic exploration of the relationship between artificial intelligence and natural perception. Invited by Audemars Piguet, she conducted field research in the Vallée de Joux, Switzerland—a journey that included factory visits, discussions, and walks through the valley landscape.

The conceptual foundation of the work draws on the ideas of technologist and philosopher Lewis Mumford, who argued in Technics and Civilization (1934) that "the key machine of the modern industrial age is not the steam engine, but the clock." The mechanical clock separated time from biological and natural cycles, transforming it into a measurable, divisible, and controllable quantity—what Mumford called industrial time. From this departure point, the artist asks: how might time be perceived by machines rather than defined by numbers?

In The Side Valley, the artist seeks to evoke what might be described as a nonlinear sense of time. The artwork imagines that a machine could be trained to re-understand meteorological time. Over a continuous period of thirty-two days, the team captured the sky every fifteen minutes from five fixed directions: east, south, west, north, and zenith, collecting a total of 6,400 images. These images served as data inputs through which the machine began to "learn the perception of natural time."

The generative core of the work is based on a deep-learning structure employing Generative Adversarial Networks (GANs). The artist divided the data into three temporal groups: morning, afternoon, and evening, then trained an independent model for each, extracting and internalizing the latent temporal correlations within the dataset. After multiple training iterations, these models were deployed within a system connected in real time to the valley’s environmental conditions, where wind speed and direction act as triggers for the generation of motion graphics. The resulting visuals are projected onto the lake surface and driven by live meteorological data, forming a feedback loop in which time, machine, and nature coexist in continuous exchange.

From this perspective, the work adopts what Agüera y Arcas calls the computational substratum as an underlying organizational force, suggesting that "meaning" is not a fixed reference but a topological generative coherence: an ecology of continuous becoming. When people gaze upon the water, time appears as a dynamic, generative phenomenon that, through real-time connection with the environment, persistently invites them to explore the unknown within the known.

*All photographic documentation and dataset were completed in the Vallée de Joux, regarded as the historical heart of Swiss watchmaking and a symbolic site where time became mechanized and precision was born.

​

​