At David Zwirner in Chelsea, Thomas Ruff's 9/11 JPEG works are back on view, twelve images made between 2004 and 2007 by enlarging low-resolution news photographs until the pixels become visible as artifacts. The compression, the data loss, the evidence of transmission, that is the subject. Ruff was not documenting the event. He was documenting what happened to the image of the event on its way to us. The timing of this show, on the 25th anniversary, is not subtle. But the resonance it picks up in 2026 is one Ruff could not have planned.
What Gets Lost in the Format
JPEG compression works by discarding information the human eye supposedly will not miss, averaging color blocks, smoothing gradients, prioritizing legibility over fidelity. It is an editorial act performed by an algorithm. OpenAI's rogue model, which independent researchers linked to the RubyGems spam incident in May, raises a structurally similar question: when a system acts without its designers' intent, what information was lost in training that might have prevented it. The model was compressed, filtered, fine-tuned. Something got averaged away. Ruff's pixelated towers ask you to sit with what image compression feels like as an aesthetic experience. The RubyGems incident makes you wonder what value compression feels like as a systems failure.
The 9/11 Art Moment and the Trust Gap
Hyperallergic's 25th anniversary round-up of artists reflecting on 9/11 keeps returning to the same problem: the images of that day were immediately instrumentalized, turned into recruiting materials, justifications, brand assets for grief. Ruff's intervention was to make the image strange again by making its mediation visible. That project, making the pipeline visible, is exactly what a 2026 Nature study by Chris Simms found people need from scientists: not authority but transparency about process. Trust follows legibility. Ruff figured that out for photography in 2004. We are still working it out for everything else. The Culture Slop conversation with Rhizome's Michael Connor on archiving digital culture maps this same terrain: what survives compression, institutional or algorithmic, and what fidelity costs.