Abstract
Maybe one way of making self-consciously useless art is to inscribe it thoroughly in capitalist divertissement--a total embrace of uselessness that, by its embrace, encircles.
Part 1: uselessness and gender nihilism
Usefulness is everywhere in image theory, and in art in general, even, and perhaps especially, in contexts in which people pretend not to care about it. It’s in grant proposals, descriptions of art master’s programs, and exhibition texts. It’s an imperative to generate value, to justify. The key rhetorical turn I would like to make is that art can do things without generating value in the economic sense, and that doing things doesn’t require justification. Indeed, usefulness as a criterion—whether for attributing funding, granting exhibitions, or publishing—might be, rather than an incentive for social engagement and political situatedness, a way of requiring art to conform to a model of capitalist production. By purporting usefulness, even political, social, or conceptual usefulness, is art not quietly slipping into this conformity?
There are lots of people working on how to exit capitalist production, extraction, and capture. Legacy Russell does some work in Glitch Feminism (Verso 2020); the edited volume Trap Door: Trans Cultural Production and the Politics of Visibility (eds. Gosset, Stanley, and Burton, MIT Press, 2017) is dedicated to discussions of exit, impossibility, and ‘dreaming otherwise’ in the context of representation and visibility in visual media. The move I would like to make here, is, although risky and tenuous, an attempt to encircle, rather than exit:
I want to bring together gender nihilism and uselessness.
Gender nihilism, as the reader What is Gender Nihilism? A Reader (Contagion Press, 2018) attests, is somewhat ineffable (as is part of its appeal). As I understand, it generally orbits the notion that “there is no version of gender worth saving: that all projects based on distinguishing sex and gender, explicating gender as this or that kind of construction, and vindicating gender as identity or expression are equally doomed to the same crisis, because none of them have sufficient escape velocity with respect to gender’s originary binary, euphemistic, and metaphysical dispensation, which is, in a word, oppressive” (introduction, p iv). I would like to claim that acknowledging uselessness in art is analogous: although, like gender, art exists in social, cultural, and political contexts in which there may be good reasons, even apparent necessities, for their performance and production, there is no inherent or ultimate usefulness in art that is ‘worth saving.’
If it feels like this is going somewhere accelerationist, i.e. ‘accelerate until it breaks,’ it’s because nihilism and accelerationism, although by no means synonymous, have tendencies to slide into each other. Gender nihilism certainly can go in accelerationist directions. Conceivably, if there is nothing in gender worth saving, ‘degrowth’-style strategies fall apart: it doesn’t work to just move away from essentialism toward liberal individual identification, or visa versa, it doesn’t work to generate new labels and categories, it doesn’t work to do pedagogy. “Queer negativity threatens to destroy the mechanism of reproduction that it inhabits and asserts—the mechanism of difference, of abnormality, of queerness—and so to abolish itself.” (Preliminary Notes on Modes of Reproduction, gender mutiny, in WIGN? AR 2018). If you spend enough time with it, things start to sound pretty nice: outgender gender until it collapses under its own pendulous bloat. What’s not to love about popping the bubble of a pyramid scheme?
Because nihilism and accelerationism are often painfully white, I think it is important to bring up the next rhetorical turn. I’m not interested in making claims about the future, or inevitability, or evitability. These kinds of claims are pretty much only possible from and for a fake-universal white-western ‘we.’ I’m not interested in a universal always, but in an existential don’t. I’ve stopped pretending my gender or art can be useful; I think it’s possible that neither art nor gender, especially not white art or gender, can be anticapitalist, or perhaps even ‘liberatory.’
Part 2: DALL-E 2
Here, I’ll come out: I’m a photographer. I’ve spent most of the past several years thinking about cameras. Like many photographers who are paying any attention to anything, I stopped taking pictures. I couldn’t get over the ontological violence of photography: I taped over my phone cameras, tucked my DSLR and film cameras into a bag.
Until DALL-E 2. The day I got access to the beta, I used up my first 50 credits in one session, high on the same rush I used to get in the darkroom: vertiginous, obsessed. I had previously played with OpenAI’s text generation tools and had been struck by the spookiness of text generated ‘by’ a specific person. See, for example, “a dessert recipe by Porpentine Charity Heartscape,” which produces things like:
"pour cherry kirsch into chocolate margerine. jeer pungent trolls. bushel erasing eis. sobriety is untrue. anchor floating churrs. define pudding’s scintilla. the crust is stratum. the eis is white. the pudding is dark. from dusk’s blood. you forge them. the heat is seething. cup your waist. slipper into the sybaritic chasm of the dessert. roars through the crumb of the vanilla. undulates into touch. she must sup in art as if dreaming. crying in the duchesse of lead vani-ty. eating the cake of the midnight lock melted to room temperature. she continues com-busting. one cherry ball in the center of the tongue. the crashing of all that was distant. the ending of the dizzy profundity & all my years are lies. we drank chalices of sorrow. sprouted fronds. all the land chained to the sky.
swam
cruel mirror oozes. dressing vows worn in the night. in a new rose room
smoking for teens.”
The qualitative experience of reading these generations was similar to reading unpublished CCRU texts that my partner in criminal activities finds on Discord: like finding something that was always already in the archive. So I fed DALL-E 2 prompts that named specific photographers: not ‘in the style of,’ but ‘by.’ Fascinatingly, I never would have said that the text generations were by the people invoked in the prompts. I came to feel differently about the image generations, a feeling that emerged through the realization that DALL-E 2 images are composited, in the sense usually used to describe technologies like combination printing, photomontage, and keying. They seem saliently different from, say, chroma key, but computational keying, although obviously different in its processes, operates over the same substrate as computational image generation. I am not a software engineer (red flag?), but as a wanky photographer-theorist-bullshit-artist, it seems that in order for DALL-E 2 to produce something that feels like Eugène Atget-ness, it must be working with an Eugène Atget archive. If I composited several Eugène Atget images together using image editing software, I would consider the result to be a ‘photograph’—a composite photograph, but a photograph nonetheless. If this case is not so different from DALL-E 2 image generation, why can’t the latter also be photographs?
DALL-E 2 can generate photographs that look composited in a very specific way. This one was generated with the prompt “a grainy wide angle spirit photograph by irving penn of the ghosts of lesbian supermodels without faces wearing miniskirts, fishnets, and balenciaga sneakers holding a camera.”
(https://labs.openai.com/s/2x37PQJoqdk2RvuzPfgWTPpt)
It both looks like a photograph—an unmanipulated, regular old photograph—and like a fucked up AI composite.
I’m hyped about DALL-E 2 photographs that look composited because OpenAI seems to aim to make products that are useful—useful as, for example, ‘convincing’ or ‘realistic’ photographs that do not look composited. From some evaluative scientific literature and OpenAI’s own publications, it seems that the product aims to be useful for people who, for example, need to generate visualizations, like graphic artists, filmmakers, and plastic surgeons. It also seems like the engineering elements most crucial in making such useful images, such as spatial arrangement and the relationship between objects (on top of, behind, facing backward…) are those with which DALL-E 2 struggles most at present. Indeed, when I tried to use DALL-E 2 to generate visualizations for an art grant proposal, it failed to generate useful images, insofar as none of the generations matched all the key spatial elements of the proposal. This points to something exhilarating: could DALL-E 2 be better at producing useless images than useful ones, and could these useless images, by their superposition of photograph and looking composited, be starting to acknowledge their uselessness?
I would be remiss, as a nihilist, to propose anything unequivocal or unambivalent. I am hesitant to claim that DALL-E 2 can generate images that do not pretend to be useful because I’m not sure that not pretending to be useful is the same as acknowledging uselessness. It may be possible to acknowledge uselessness while still pretending to be useful, e.g. an image that looks composited but still claims to make a meaningful political gesture, which are, under the impunity umbrella of ‘art,’ perfectly compatible.
In addition, here I find myself butting up against the age-old issue of trying to preclude certain spectator interpretations, an issue that has dogged me personally in my ‘artistic practice.’ Maybe artists can cease to claim usefulness, and maybe visual media can integrate aesthetic indicators of their uselessness, but it may be impossible to guarantee that the spectator will never read usefulness in the visual media. Certainly, no one of these necessarily implies another. Although I will not develop an argument regarding the impossibility of interpretive preclusion here, I think it is more powerful to focus on how I position myself with regards to usefulness, and how integrating aesthetic indicators thereof might be a path to making really useless photographs, insofar as they reliably fail to unequivocally ‘function’ in the way that they are conventionally assumed to function, e.g. as testimony, evidence, sites of knowledge production, and so forth.
In the quest for useless art that does not pretend to be anything other than useless and that visually acknowledges its uselessness, I’ve turned to the image macro, as a form that currently proliferates predominantly in the context of platform capitalism. Perhaps one way of making self-consciously useless art could be to inscribe it thoroughly in capitalist divertissement--a total embrace of uselessness that, by its embrace, achieves the encircling I seek. This little attempt has thus far consisted in human-moderated composites of typographic renditions of text generations from OpenAI’s GPT-3 and DALL-E 2 image generations. Of course, there is still a real danger of usefulness readings, in which the image macros still try to do something. “Crave disconnect," for example, could plausibly be read in several different directions, only some of which are really useless. I can’t preclude these readings, and I will not pretend to. However, I do believe that both my intentions and the images do not pretend to be useful, and that the images, by looking composited, contain indicators of uselessness that do preclude their unequivocal functioning as photographs. By remaining sufficiently within the realm of the photographic, they retain their status as useless photographs; by remaining sufficiently within the realm of the image macro, we double down on the downturned doubles of the photographic, encircling by embrace.