When a neural network draws a clock, the hands most often stop at 10:10.
March 1, 2025
This is not a point in time, but a stable visual motif that repeats for no reason. In this work I explore what resists digital interpretation: what can be digitized, and what always slips away.
Curator: Murad Khalikov — @mooreeque
Documentation: Lola Graham — @lolagrhm
additional link -
https://lolagrhm.wfolio.pro/disk/10-10
Documentation: Lola Graham — @lolagrhm
additional link -
https://lolagrhm.wfolio.pro/disk/10-10
The room was mapped with laser levels, yet chance remained — a randomly chosen square became the core element. Its image was given to a neural network, which first described it in words and then generated versions that were almost identical, but slightly distorted. One of these versions was suspended in space, aligning perfectly with the laser grid and creating a subtle shift in the perception of reality. The same text was turned into sound and sent through a vibration speaker, turning a cabinet into a sound source. An algorithm transformed the description of the space into a rhythmic recitative, with words adapting to the structure and following a set tempo.
But one element in the space resisted digital reconstruction — smoke. It rose from a pipe, though not a stove pipe — it was produced by a smoke machine. Crossing the laser grid, it shattered along its coordinates, creating unrepeatable patterns. Without heat, without smell, without traces, it existed only in the moment — pure appearance without substance, just like “10:10” — a moment when reality and its digital double coincide, yet never become the same.