What Survives the Algorithm

The Thumbnail Problem

Every serious work of art, in the current environment, must negotiate the thumbnail. The painting, the sculpture, the photograph — reduced, in its first encounter with most of its viewers, to a rectangle approximately one centimeter by one centimeter, encountered on a phone screen while someone waits for a coffee or sits on a bus or scrolls through a feed in the late-night loneliness of a week they do not want to think about.

This is the condition of art now. It has always been true that most people encounter most art as reproduction rather than original. The art history slide, the postcard in the museum shop, the coffee-table book. But the thumbnail is a new kind of reproduction: not an attempt to convey the work with some fidelity, but a bid for attention in a competition for attention, against ten thousand other thumbnails, lasting approximately 1.7 seconds before a decision is made.

What survives this? What does the algorithm — in the sense of both the literal machine-learning systems that rank and surface images and the broader attentional economy they instantiate — allow through, and what does it suppress?

The Dimensional Problem

Large-scale painting does not thumbnail well. This is nearly definitional: a work that derives its meaning partly from its scale — that requires you to stand before it and be enveloped by it, to feel its weight in relation to your body — cannot be adequately conveyed in a rectangle smaller than your thumbnail. Mark Rothko understood this. The Color Field painters generally understood this. You cannot feel the pressure of a Barnett Newman Zip on your retina when you encounter it on an Instagram grid.

The work that resists the thumbnail is the work that is still doing something that thumbnails cannot contain.

This is, I want to argue, not a complaint about platforms. It is a description of a condition, and the condition has a productive corollary: if a work survives reduction, it tells you something. If a work is completely legible in thumbnail — if the thumbnail conveys everything the work does — that also tells you something.

The work that resists the thumbnail is the work that is still doing something that thumbnails cannot contain.

The paintings that photograph particularly well are, not coincidentally, often the paintings that function most like graphic design: high contrast, strong composition, legible iconography, immediate visual impact. None of these are bad qualities. But they are qualities of a particular kind of art — art that works at the level of the image rather than the object, art that is complete in its appearance rather than its presence.

Time-Based Resistance

Some works resist the algorithm not through scale but through time. The video installation that requires forty-five minutes to unfold, the sound work that only coheres after ten minutes of listening, the sculpture that changes under different light conditions over the course of a day — these works cannot be adequately documented. They are, by design or by accident, immune to the conditions of the feed.

This immunity is, I think, increasingly precious. The culture is producing, at scale, objects that work at the speed of the feed. Fast read, fast impact, fast move. The artwork that insists on its own temporality — that will not give itself up to a three-second encounter — is holding a position against the dominant logic of distribution.

I am not saying that difficulty equals virtue. Plenty of difficult art is simply obscure or lazy. But the works that have occupied me most intensely in the last decade of criticism are almost uniformly works that gave themselves up slowly. Agnes Martin. On Kawara. Tacita Dean's film works. Theaster Gates at his most monumental. These are artists whose work unfolds at its own pace, and that pace is not the pace of the feed.

What Gets Made in Response

The algorithm does not only shape what gets seen. It shapes what gets made. This is the subtler and more important claim.

Artists, like everyone else, exist within their distribution environment. The artists of the fifteenth century made works for the conditions of the church and the private palace. The artists of the nineteenth century made works for the salon and the public exhibition. The artists of the late twentieth century made works for the white cube and its international circuit of art fairs and biennials. And now artists are making works for the image economy — for the conditions of visibility that the platforms define.

Some of this is conscious and sophisticated: artists who engage with the platform's logic as their subject, who make work about visibility, virality, image circulation. This is legitimate and interesting. But some of it is less deliberate: a gravitational pull toward the immediately legible, the shareable, the aesthetically pleasing in a way that reproduces well. A slow pressure toward the work that works as an image.

The question for a critic is: what is being lost? What kinds of work are becoming harder to make, or harder to sustain, or harder to have anyone take seriously, because they resist the conditions of distribution? I think about this often, and with increasing urgency.

Attention as the Last Medium

Here is the argument I keep arriving at: attention is the last medium. Not canvas, not stone, not light, not duration. The medium in which all other media exist, for a viewer, is the quality of attention they bring. And the quality of attention available to most viewers — trained, over years, by the rhythms of the feed — is not the quality of attention most serious art requires.

This is a crisis that the art world is only beginning to name. It is not a crisis of quality — there are extraordinary artists working now, in every medium. It is a crisis of reception: the frame within which art is encountered, the duration for which it can be held, the kind of looking that viewers bring.

The art that will survive the algorithm is the art that creates its own conditions of attention — that insists, through its structure or its scale or its pace, on a different relationship with time. This art will not be the most visible. It will not accumulate the followers or the resale value or the name recognition. But it will be doing what art has always done at its best: making an unreasonable demand on your presence, and rewarding that presence with something that cannot be obtained any other way.