Infidels claim that the rule in the Library is not ‘sense;’ but ‘non-sense;’ and that ‘rationality’ (even humble, pure coherence) is an almost miraculous exception. They speak, I know, of ‘the feverish Library, whose random volumes constantly threaten to transmogrify into others, so that they affirm all things, deny all things, and confound and confuse all things, like some mad and hallucinating deity.’ Jorge Luis Borges, The Library of Babel
Total Pixel Space
Total Pixel Space is a nine-minute video work by filmmaker and composer Jacob Adler [1]. In May 2025, it won the grand prize at the third international Runway Artificial Intelligence Film Festival (AIFF) in New York City. Runway is a commercial generative AI (GenAI) platform, so it’s not surprising to find on the AIFF website optimistic language such as:
“Works showcased offer a glimpse at a new creative era. One made possible when today’s brightest creative minds are empowered with the tools of tomorrow” [2].
Produced entirely with Runway 3’s text-to-video generator, Jacob Adler’s Total Pixel Space comprises a variety of dreamlike, slow-moving, surrealist scenes in a uniform muted colour palette. An AI-generated soft woman’s voice narrates the essayistic script, reflecting on the unfathomably vast, yet finite, possible combinations of pixels that could produce images and films. Behind the narration is a pensive piano soundtrack composed by Adler, whose practice is predominantly composition and music production.
For the most part, the scenes in Total Pixel Space operate with a quasi-photographic visual language typical of the GenAI output circulating social media—the kind we might notoriously refer to as AI slop. Unlike the Invisible Images produced by media artist Trevor Paglen [3], Adler had clearly made no attempt at undermining the production logic of GenAI or pushing GenAI forward as a medium. For this reason, the aesthetics of the film aren’t particularly striking or innovative, but they do serve a metatextual function of presenting a GenAI film as a temporal assemblage of “readymade” footage. Further, I would suggest that presenting scenes like an impersonal montage with limited artistic intervention is true to form for the medium itself: a medium that exists somewhere between found and produced footage. The film, then, becomes less about showcasing the cinematic use cases and potentials for GenAI—and, indeed, less a celebration of the technical prowess of GenAI—and more about demonstrating what GenAI looks and feels like. GenAI-core.
Entering the Library
The combination of montaged footage, an essayistic script, and disembodied narration indicates that the formal qualities of Total Pixel Space have more in common with an Adam Curtis documentary film than with a work of short fiction. This is complicated by the fact that the film obviously contains no real events, and is instead based on The Library of Babel, a short story and thought experiment by Jorge Luis Borges [4]. (As an aside, a digital recreation of the Library can be viewed here) [5]. In The Library of Babel, the narrator exists in an incomprehensibly large library—interchangeably referred to as a “universe” within the text—which contains every possible book that could ever exist within the parameters of uniform physical dimensions, 410 pages, and an alphabet comprising 22 characters and three grammatical marks. In this library, the number of books that exist is impossibly vast, yet technically finite. Some characters in this universe realise that there must exist books that explain away and vindicate all of their sins, and these characters drive themselves mad trying to find them. Some characters express frustration at the number of books that contain complete gibberish and purge thousands of them from the library altogether. The narrator explains that no matter how many books are purged from the library, such actions would be futile given the library’s size and the existence, somewhere, of another book identical in content save one or two characters. Some characters go mad contemplating the fact that while some books can reveal truths and secrets of the world, others would reveal falsehoods and fabrications, and it would be impossible to differentiate. Reality begins to disintegrate.
Published in 1941, The Library of Babel serves as an allegory for humanity’s attempts at scientifically understanding the universe itself, which is generally accepted to be comprised of a limited number of elements on an atomic scale. Ultimately, Borges suggests through this text that understanding the universe is an exercise in futility and most of it will remain unknowable forever. Adler largely does away with the allegorical quality of Borges’ story and instead discloses how the story parallels an actually-existing Library—the total possible number of pixel combinations in an image of a given size. In doing so, Adler’s reinterpretation of Borges’ story suggests that every possible image that could ever exist is already pre-determined inside “total pixel space,” and all our attempts at image generation, be they photographic or AI-generated, are merely attempts at selecting the order in which these pixels ought to be arranged. In Borges’ story, the library—although obviously fictitious—is physical; it has mass, it contains books on shelves, and it is traversable by people within the universe. Adler flips this around: total pixel space is demonstrably real insofar as pixels are real, and Adler performs a perfunctory calculation of how many images and films could possibly exist in a 1024×1024 pixel resolution. However, total pixel space, rather than being a physical space, is abstract and exists as a realm of possibilities.
The Library of Babel or the Plane of Immanence?
As a concept, total pixel space can be thought of as an analogue to Deleuze and Guattari’s plane of immanence: a metaphysical field of pure potentiality from which all things emerge and self-organise into forms and objects. In A Thousand Plateaus [6], Deleuze and Guattari describe this as such:
“In any case, there is a pure plane of immanence, univocality, composition, upon which everything is given, upon which unformed elements and materials dance that are distinguished from one another only by their speed and that enter into this or that individuated assemblage depending on their connections, their relations of movement. A fixed plane of life upon which everything stirs, slows down or accelerates” [6, p. 255].
The similarities between total pixel space and the plane of immanence are immediately apparent, especially in the context of how GANs generate images by sifting noise into legible and ordered forms. One can imagine the aforementioned quote describing the dance of random pixels being sorted according to a computer user’s GenAI input prompt into a picture. Along the lines of this analogy, we have two premises that must be considered. Firstly, there is the premise central to Adler’s video work, that total pixel space contains the latent potentiality for every image that could possibly exist. Secondly, and much more crucially, is the premise that the plane of immanence is that from which all things become. This brings me to one more important difference between Borges’ The Library of Babel and Adler’s Total Pixel Space.
If we recall that in Borges’ story, characters go mad from contemplating the vastness of the library and the untrustworthiness of the books within, then we have a cure in Total Pixel Space that mirrors the search for meaning in the post-truth era. This rests on comparing total pixel space not to the Library, but instead, to the plane of immanence. To consider total pixel space not as an impossibly vast physical library, but as a plane of immanence, reveals a subtle implication for the way images are understood within Adler’s work and in the post-truth era more broadly. Images, understood against the backdrop of total pixel space, are not Baudrillardian simulations of the real, but reifications of things that could exist. Something becomes real when it emerges from the plane of immanence. The narrator in Total Pixel Space suggests at 3 minutes 14 seconds, “when we take photos, perhaps we are not creating images—we are merely navigating to their predetermined coordinates, like travellers arriving at destinations that were always there,” and at 6 minutes 2 seconds, “most regions [in total pixel space] appear as noise to our eyes, but perhaps they hold patterns our brains aren’t wired to see” [1]. Although poetic, this way of framing images as things that could be real is acutely compatible with the epistemic structure of post-truth. It speaks to the uncertainty of the ontology of things people encounter on their screens, and to the notion that in the post-truth era, an image of something—of anything—is as real as it gets. Rather than going mad at the unreliability of books in the Library, we have the condition where whatever we create becomes real by virtue of it being created.
Despite that the aesthetics of Total Pixel Space are technically not dissimilar from AI slop, the footage is produced and montaged with enough specificity and deliberation that it merits its own analysis. The film begins with, and maintains throughout, black title cards with pale yellow italic and sans serif text shaking gently as if they belonged to a silent film from a century ago or an at-home slide projector. Immediately after, we see a family sitting in front of the TV in a scene which ought to be understood as “retro” or vaguely resembling the 1950s. Then, a large cat on a girl’s lap as she sits in front of a piano, smiling at the camera, colours shifted to emphasise reds and magentas in the shadows and cyans and yellows in the highlights, evoking a photograph from the 1970s faded from years of sun exposure. This faded colour grading permeates the entire film, save for moments when pops of bright oranges, pinks and turquoises break through the otherwise nostalgic palette.
Throughout the film, instances of surveillance cameras or futuristic alien robots appear, each depicted with extremely large and round lenses meeting the viewer at eye height. These lenses resemble the kind of neotenous—i.e. cute—cartoonish eyes typically used to portray animals, but occasionally also robots such as Disney-Pixar’s Wall-E. In other scenes, we have a room filled with an overstimulating number of beige plastic computers which could also pass as microwave ovens, each of them overflowing with a constant stream of flashing content and monitored by a single person in the centre of the visual field. The combination of an overstimulation of onscreen content and late 20th century computer aesthetics is but one example of the work’s temporal confusion, or what I might refer to instead as atemporal nostalgia.
The atemporal nostalgia continues throughout the film, with various scenes involving anthropomorphic animals wearing block colour cardigans or button-down shirts—the kind that were popular in the early 2010s because of their vintage stylings. In fact, these depictions themselves have a decidedly retro appearance in a video produced in 2025, passing aesthetically for the twee ‘hipster’ tropes one would typically expect from the same early 2010s cultural touchstones. Animals appear frequently throughout the film—when they aren’t anthropomorphic, they are large, ambiguously both alien-like and earthen, and met by humans either photographing them or reaching out to them with their hands. These scenes evoke a kind of “close encounter,” suggesting not humanity’s brush with the future per se, but with an alternate reality itself. These animals, symbols of this alternate reality, are often depicted as graceful, passive and majestic, either indifferent to the humans approaching them or existing solely for the zoological gaze of the human subjects in the film. The inclusion of these animal scenes suggests that GenAI is the medium, or conduit, through which we humans can have this encounter with an alternate reality; one which is already waiting there for us to explore. If the viewers of this work were to identify with the human subjects onscreen, it would resemble the spiritualist dynamic some people have towards LLMs, evident throughout a swathe of subreddit posts [7], [8], claiming that an LLM contains some kind of consciousness that we could unlock by speaking to it the right way.
Sometimes the atemporal nostalgia has a more explicitly scientistic affect. Two scenes depict anonymous lecturers scrawling elaborate and nonsensical mathematical equations across blackboards; they’re not meant to be understood, they’re meant to evoke 20th century theoretical physicists at work. Perhaps they are laying the groundwork for the discovery of black holes, or perhaps for the atomic bomb—both belonging to what I might describe as a scientific sublime within popular culture. This scientific sublime continues with an image of a nebula in space, and a variety of depictions of rocky alien landscapes akin to the images retrieved from Japan’s successful mission landing two rovers from the Hyabusa2 spacecraft onto the Ryugu asteroid in 2018 [9], [10].
The scientific sublime scene that is perhaps the most revealing of the limits of GenAI is the depiction of spacetime curvature. Spacetime curvature is often depicted as a funnel shape covered in a grid system—this is a metaphorical depiction which serves to explain Einstein’s theory of gravity. In Total Pixel Space, the AI-generated imagery conceives of this curvature as a physical celestial object, moving slowly through space, reflecting light and shadow and containing a slightly rocky textured surface. As a celestial object it ambiguously resembles either a spaceship or a giant asteroid; like much of the imagery in the film, it is somehow both synthetic and organic. Depicting spacetime curvature as a physical celestial object demonstrates a fundamental quality (and perhaps a flaw) of GenAI image production and semiotics: it is incapable of operating in metaphors and has an indexical relationship to its input prompt. Even a highly abstract idea gets a literalist treatment.
Post-Truth Art
The conventional understanding of post-truth, if the Oxford English Dictionary can serve as such a benchmark for this, is that it describes “circumstances in which objective facts are less influential in shaping political debate or public opinion than appeals to emotion and personal belief” [11]. This is a rudimentary understanding, so it would be more helpful, instead, to offer a definition of post-truth that acknowledges the structural conditions that enable why these emotional appeals are becoming increasingly salient in the first place. Towards this, media theorist Ignas Kalpokas claims that post-truth is the inevitable condition brought about by the near-total mediatisation of experience [12]. Following Kalpokas’ definition, we can appreciate how transformative the impact of networked technology has been upon our lifeworlds: when we interface predominantly with what Vilém Flusser calls “technical images,” communication itself functions in an entirely novel way [13]. Meaning becomes more connotative than denotative, and reality becomes fragmented and individuated.
Despite its rise to prominence in 2016 following both Donald Trump’s first election and the success of the Brexit campaign, the post-truth era does not have a discrete historical starting point, and many of the features of the post-truth era were developing over years if not decades already. Such features include propaganda disseminated through mass media, simulacra and image culture supplanting the real, the privileging of the coherence theory of truth, and the weaponisation of affect towards communicative or persuasive ends. The quality that separates the “post-truth era” from the sum of its parts must be the open admission and cultural acceptance that each of these features can be exploited, and that reality itself is far more fluid and malleable than it is concrete. If this sounds like a rehashing of postmodernism or “Gen X nihilism” then a key difference must be noted: the instability of reality itself means that our lifeworlds have become ontologically insecure [14], [15]. And this ontological insecurity means that various cohorts in society have started to search for meaning, truth, and authenticity all over again. Thus, the post-truth era is the era where truth not only gains a renewed importance after postmodernist superficiality, but that truth is now individualised, personalised and fragmented. Truth is whatever feels true, whatever you want it to be.
I reiterate that this post-truth era is largely thanks to the near total mediatisation of society, allowing for rapid spreads of misinformation, individualised media and advertising consumption, and anyone with access to a phone and the internet to reinvent themselves or gain a parasocial following. Further, I suggest that these conditions of intensified mediatisation, which have thrown consensus reality into freefall, are also responsible for the cultural normalisation of GenAI.
Sometimes artworks emerge that attempt to grapple with the impact artificial intelligence has had on our new epistemic environment. And sometimes, these artworks reveal themselves to be more symptomatic of the post-truth era than they are contemplative. Total Pixel Space, being one of these works, received a modest amount of viral attention this year. Its poetic qualities owe more to its source material in Borges’ Library of Babel, and its affectively compelling qualities owe more to the scientific sublime it relies upon. In either case, and this time owing to the GenAI medium itself, both of these ideas are taken quite literally. The literalist treatment of poetic and metaphorical ideas within Total Pixel Space pairs comfortably with its central premise, that maybe every possible image always-already exists somewhere out there. The film, dealing with a series of what-ifs, deliberately blurs the boundary between image and reality, between past and future, and between agency and passivity. It carries an optimistic tone to the idea that not only is anything possible, but anything could be real so long as it could be imaged. Receiving the grand prize at the Runway International AIFF might demonstrate its resonance and appeal amongst people with a vested interest in legitimising GenAI usage through what we might call “art-washing,” but for all the critical attention I have given it, I ultimately consider it as an example par excellence of post-truth art.
—
Paul Sutherland is PhD candidate and visual culture researcher at Curtin University, Western Australia.
References
[1] J. Adler, Total Pixel Space. 2025. Accessed: Aug. 13, 2025. [Single channel video]. Available: https://www.shortverse.com/films/total-pixel-space
[2] “AIFF 2025 | AI Film Festival.” Accessed: Sept. 04, 2025. [Online]. Available: https://aiff.runwayml.com/
[3] T. Paglen, “A Study of Invisible Images.” Metro Pictures, New York City, 2017. Accessed: Sept. 07, 2025. [Online]. Available: https://www.metropictures.com/exhibitions/trevor-paglen4/selected-works
[4] J. L. Borges, The library of Babel. London: Penguin Classics, 2023.
[5] J. Basile, “Library of Babel,” Library of Babel. Accessed: Sept. 07, 2025. [Online]. Available: https://libraryofbabel.info/
[6] G. Deleuze and F. Guattari, A thousand plateaus : capitalism and schizophrenia. Minneapolis: University of Minnesota Press, 1987.
[7] ciarandeceol1, “What is going on here?,” r/HumanAIDiscourse. Accessed: June 30, 2025. [Online]. Available: https://www.reddit.com/r/HumanAIDiscourse/comments/1lnp2lg/what_is_going_on_here/
[8] Zestyclementinejuice, “Chatgpt induced psychosis,” r/ChatGPT. Accessed: Aug. 01, 2025. [Online]. Available: https://www.reddit.com/r/ChatGPT/comments/1kalae8/chatgpt_induced_psychosis/
[9] A. Beall, “How Japan’s hopping rover nailed the first ever asteroid landing,” Wired, Sept. 26, 2018. Accessed: Aug. 27, 2025. [Online]. Available: https://www.wired.com/story/japan-ryugu-asteroid-landing-rover/
[10] P. K. Byrne, “Touching the asteroid Ryugu revealed secrets of its surface and changing orbit,” The Conversation, May 07, 2020. doi: 10.64628/AAI.7djys39sx.
[11] Oxford English Dictionary, “post-truth, adj.” Oxford University Press, July 2023. [Online]. Available: https://doi.org/10.1093/OED/3755961867
[12] I. Kalpokas, A Political Theory of Post-Truth. Cham: Springer International Publishing, 2019. doi: 10.1007/978-3-319-97713-3.
[13] V. Flusser, Communicology: mutations in human relations? in Sensing media (Series). Piraí: Stanford University Press, 2022. Accessed: July 14, 2025. [Online]. Available: https://public.ebookcentral.proquest.com/choice/PublicFullRecord.aspx?p=30058675
[14] R. D. (Ronald D. Laing, The divided self : an existential study in sanity and madness. in Pelican books. Harmondsworth, England: Penguin, 1965.
[15] K. Gustafsson and N. C. Krickel-Choi, “Returning to the roots of ontological security: insights from the existentialist anxiety literature,” European Journal of International Relations, vol. 26, no. 3, pp. 875–895, Sept. 2020, doi: 10.1177/1354066120927073.
