Domestic Chronotopes

Artist: Roger Weiss

Title: Domestic Chronotopes: Archaeology of the Everyday
Multimedia installation, variable dimensions, 2024–2025.

AI specifically developed by Brandcraft

Link to the installation:

Domestic Chronotopes: Archaeology of the Everyday is a hybrid installation that explores the subtle threshold where the ordinary transcends its apparent banality. It begins with a simple question: what happens to the gestures we perform without noticing? If no one witnesses them, not even ourselves, have they truly happened?

The installation inhabits a reduced domestic landscape composed of geometric volumes, built at a 1:1 scale. These parallelepipeds recall the forms of a table, a carpet, a toilet, and other elements of everyday life. Though inert, they come to life through embedded screens that loop recordings of common actions associated with these objects: eating, cleaning, washing.

Two custom-designed artificial intelligences intervene on distinct levels. The first manipulates the temporality of the videos in response to external data. Each work draws from a specific archive: two thousand years of conflict, a century of financial fluctuations, demographic birth trends, or the evolution of the human–machine relationship from the early twentieth century to a speculative future. These data are translated into an intensity scale that alters the rhythm of the gesture. The result reveals how macro forces infiltrate even the most intimate actions, rendering Bachelard's "first universe" no longer a sufficient refuge.

The second AI removes the mediating tools of each action (cultural prosthetics like cutlery or vacuum cleaners) transforming purposeful movement into abstract, functionless sequences.

Through this double subtraction, time and matter dissolve. The gesture emerges in its bare form, as a choreography suspended between memory and automatism. The familiar becomes strange again.

Two custom-designed artificial intelligences were developed in collaboration with Roger Weiss for the project:

• The first AI modulates the temporality of gestures based on historical datasets (conflicts, financial fluctuations, demographic trends, pandemics, and the human–machine relationship) using Python with Torchaudio and Pandas.
• The second AI detects and removes the objects mediating the action (e.g. cutlery, vacuum cleaner) through deep-learning video inpainting techniques specifically developed with PyTorch, OpenCV, and ffmpeg.