2026–ongoing

YOU DO NOT CONTROL THIS MACHINE

An artwork in which an institution contributes bounded computational capacity while relinquishing authority over the purposes enacted through it.

Computational infrastructure, distributed execution, installation and institutional protocol

Concept rendering of a dark gallery installation with computational infrastructure beside a luminous field of abstract visual residue.
Concept rendering for the developing installation.

YOU DO NOT CONTROL THIS MACHINE asks an institution to contribute computational capacity while surrendering a narrow authority over its use. An external requester declares a resource envelope. A compatible worker commits that capacity before receiving the encrypted execution payload. The requester retains the consequential result. The institution keeps the machine, its operational obligations, and a visible residue produced from privacy-reduced traces of the computation.

The host defines the powers available to a workload: isolation, quotas, networking, maintenance and legal limits. Within those limits, it does not select what an otherwise-conforming workload is for. The work separates ownership of infrastructure from semantic authority over every permissible use.

The current prototype schedules and executes real CPU workloads, coordinates multiple workers and returns encrypted results to the requester. Its institutional form remains under development. The artwork requires an independently administered host willing to contribute real capacity under the same division of authority.

A beautiful exhaust

The useful output never appears in the installation. Completed jobs deposit anonymous visual matter into a persistent field; live activity folds and displaces it; older activity sediments and decays. The gallery sees formal consequences of computation whose purpose and result remain outside its possession.

The present renderer is an initial state. The installation is being developed through research into relief, colour, light, physical depth, architectural placement and changing perception. Its ecological and material account is part of the work: greater resource consumption must not become automatically equivalent to greater visual intensity.

Current research

Development centres on four connected problems: independent institutional stewardship; operational security for requester-supplied workloads; public access that does not privilege only technically expert users; and a spatial manifestation whose form is materially caused by external use without exposing requester content.

The project extends artistic histories of chance, instruction, delegated execution and participation by relocating relinquishment from authorship or form to infrastructural purpose. Construction remains necessary because the division of authority has to survive administration rather than remain an intention.