Speech at INC Exit Fest, OT 301, Amsterdam, 25th June, 2026.
Online video has become the general condition of social media, a streaming vortex driven by extractivism and behavioural targeting. Reflecting on my experience as a member of CDI-TV,[1] a collective contributing to StreamArtNetwork (SAN), I want to consider this condition from a specific angle, what I am calling poor chroma: the shimmering, unstable borders that appear where chroma-keyed bodies fail to seamlessly converge with the background, where a slight glitched-out halo forms around hair and fingers, the signs of a spill that doesn’t quite wash out. They might seem like minor technical imperfections, but I argue that poor chroma offers one way to understand the stakes of streaming in terms of tactical publishing, strategic power and the logistical operations of streaming infrastructures. Poor chroma can help us understand how liveness, distribution and collectivity are both held together and apart by corporate stacks; in other words, how they are interfaced and, therefore, how the mediation within the vortex becomes conceivable.[2]
THE UNSTABLE OBJECT
Let’s begin with a classic premise of video theory: the fact that video has never had a stable definition as a medium. As Sean Cubitt put it: “there is no essential form of video.”[3] Video’s impurity was inseparable from its potential, both actualised through multiple, contradictory uses, including surveillance, art and performance, television, home recording, feedback systems, activist media, webcams, tutorials, gaming, videocalls and streams. Video has always been plural, hybrid, difficult to contain; qualities that passed on to tactical media from the very first Next Five Minutes on tactical television.
There is one thread, however, which I want to pick up on: instantaneity. The live image fascinated early video theory, particularly in the forms of ‘the monitor’ and feedback, through which images could be returned quickly enough to alter the situations from which they emerged. This is a lineage that runs from Paul Ryan to Vilém Flusser, from Dan Graham to Rosalind Krauss,[4] where reflexive instantaneity might mean either cybernetic becoming or a new form of narcissism, along with the promise of talking back to television, a critical ambition conveyed in Ryan’s slogan: ‘VT is not TV’ (videotape is not television).[5]
In the 1990s, Maurizio Lazzarato gave this predilection onto-political force in his Deleuzean account of video as a technology of time rather than movement. Drawing inspiration from Nam June Paik’s observation that “video imitates not nature, but time,”[6] Lazzarato foregrounded video’s live quality as a capacity to intervene in processes of collective subjectivation (‘machines to crystallise time’). Crucially, he also warned that this capacity can be captured: television, for him, was an apparatus of power constituted as the denial of video’s real time. When read alongside his later development of these ideas in Signs and Machines,[7] the trajectory becomes clear: the platform is now the apparatus that enacts this capture. Livestreaming and online video are delivered through complex time-managed assemblages, involving latency settings, buffers, codecs, adaptive bitrate ladders and many other obscure mechanisms. The old promise of video was instantaneity, and the platform stack today administers it.
TEMPORAL BANKS
The livestream thus flows into the platform as into a reservoir, as Michel Serres might put it. In Hermes IV: Distribution, in an essay on the history of motors (moteurs), Serres argues that information technologies function by perpetuating differences, but only from a filled reservoir.[8] He condenses the schema, relevant to theories of technics after the industrial revolution, as: “Réservoir, Différence, Circulation” – a triad that can also assist in conceptualising the platform logic of liveness. Within this schema, the stream becomes a reserve of time, affect and social charge from which further differences can be generated (as we might say after LLMs), ranging from clips to reactions, conversion and retention curves, even training data. Circulation then puts those differences to work through transmissions and feeds in pursuit of profit. In the subrepresentational sense of datafication, the stream is already archival before it goes live. It moves from capture to liveness, followed by replays and remixed content. The influence of reality TV’s production logic is apparent here, epitomised by the ‘content house’ or ‘collab house,’ where sponsored influencers enact a kind of primitive accumulation. In the extreme, everything is streamed, chronically captured and banked for subsequent circulation. But this does not mean liveness has lost all force. The live moment still has an intensity: a risk, a danger, a vulnerability, sometimes just awkwardness. The key question is how this instantaneity survives when it flows so quickly into architectures of extraction; whether the reservoir can still pool and release a shared or collective time, rather than simply feeding the platform’s channels of circulation.
THE DIGITAL BONFIRE
On this point, for me, the work of UKRAiNATV is a deep inspiration. Following Russia’s full-scale invasion, they would light what they called a ‘digital bonfire:’ a chroma-keyed livestream often sent from StreamArtStudio in Kraków across dispersed geographies, and around which a core collective would hang out, discuss, perform, DJ, cook, sometimes even sleep on stream (all streaming ultimately takes place in a bedroom). Their practice emerged from conditions of war and crisis, where ordinary forms of gathering, care and co-existence have been violently interrupted. The solidarity concept they developed through this practice is known as hybrid togetherness.[9]
StreamArtNetwork performance at Transmediale 2026, Berlin.
Central to UKRAiNATV’s practice is a refusal to respond to violence by reproducing war as spectacle; instead, they create an ambient counter-flow, what Geert Lovink has described as an “oasis of slow tech.”[10] Their aesthetics often involve the multiplication of frames, extensive use of feedback loops, and, of course, the chroma-key itself; all of which maintains a shared, distributed spatio-temporal field in conditions of extreme pressure. Against platform liveness as an extractive reservoir, UKRAiNATV retrieves instantaneity by tending the bonfire; in doing so, they also accumulate traces: the frayed edges and shimmering feedback of poor chroma. Their streams, known as EFIR, are a way of keeping something open to a collective gathering, including the possibility of return. They show that unique forms of sociality, even culture, can still be produced inside the heavily formatted conditions of the techno-social.[11] And all this remains possible even when those conditions are hostile or deeply compromised.
VIDEO AS PUBLISHING
At CDI-TV, we became interested in StreamArt aesthetics because of the way they explicitly take up live video as a critical composition, rather than an afterthought. A great deal of institutional streaming is still determined by the auxiliary logics of recorded talks, online meetings, remote teaching, maybe the odd promotional clip. The pandemic period looms over all this as the period where online video was widely rolled out, but mainly as a fallback infrastructure. CDI-TV counters these dynamics by treating video foremost as publishing, taking inspiration from research projects like Expanded Publishing at INC by emphasizing public-making and deliberate editorial distribution.[12] A few months ago, we published our transcripts from Season One as a downloadable book, Hybrid Encounters.[13] We’re interested in video first as a mode of thought: a milieu of experience, individuation or subjectivation. Radical video pedagogy should not mean an unreconstructed lecture or Zoom seminar.
Poor chroma embodies this mindset, refusing any fantasy of transparent playback. A good chroma-key might make the green screen (almost) disappear, separating foreground from background so smoothly that the operation vanishes, but StreamArt leaves the seams perceptible, sensible: whether through feedback, the multiplication of framing or visual spillage that marks the persistence of borders awash in a field of calculation. Bodies appear processed. Poor chroma updates what Hito Steyerl called ‘the poor image,’ but within the context of live compositing. For Steyerl, the poor image is “a copy in motion” that loses resolution as it travels, trading fidelity for speed, circulation and social life.[14] Poor chroma is slightly different, however: it trades compatibility with platform technics and the influencer economy for a sociality of care. Poor chroma foregrounds the role of green as a sacrificial colour, a persistent reminder of the costs and exclusions of the livestream image.
RESISTANCE AND ITS LIMITS
Video supports counter-publics, but livestreaming does not automatically mean critical resistance. Livestreams can document abuse and sustain solidarity, but they can also intensify harassment, manufacture far-right spectacle and monetise outrage. When organised networks are easily absorbed by platforms and their energies are metabolised as engagement, a critical approach should reveal or reorganise the conditions of circulation, which is to say, the logistical worlds at stake.[15] Dealing with an erosion of autonomy and a hostile infrastructure is a difficult problem. How do you gather together without becoming merely an audience segment? How do you keep something alive without succumbing to the audiencing of the influencer? StreamArt is, for now, a contradictory formation: there are increasing experiments with PeerTube and sites like The Inbetween, but it often moves through YouTube and Twitch. The stream reaches us because it passes through Big Tech data centres and their systems, one way or another. The digital bonfire burns inside the corporate stack. However, StreamArt treats that stack as a problematic field and doesn’t mistake access for empowerment. It projects hope from within conditions where the platform formats the future.
VIDEO WITHOUT VIDEO
The platform’s formatting needs to be tackled at the infrastructural level. This is what THE VOID’s formulation of ‘video without video’ suggests: that is, video understood as technical media. Software studies can be of use by extending analysis to the non-visual dimension of compression algorithms, network protocols, playback devices and other techno-settlements of power, in the spirit of what Harun Farocki called operational images.[16] Poor chroma becomes a diagnostic force here, accumulating traces from interfaces that continually assemble images of bodies through signals. Chroma-key filters, for instance, work through colour distance and luminance, while codecs predict blocks and motion as the stream enters the platform; these operations pull in different directions and both contribute to the indeterminacy of StreamArt aesthetics. As hybrid transmissions leave the platform, moreover, they flow through adaptive bitrate ladders, where content from the same event is multiplied into variants of different resolutions. One viewer receives a crisp high-def image, another receives a more compressed, blurry or blocky experience. The ladder preserves connectivity, dividing togetherness by differentiating it. The event, in other words, survives through graceful degradation, designed to keep people with different bandwidths and devices connected by turning volatility and inequality into individualised losses. Within laddering is a micro-temporal regime of resilience which strives to deliver more of the same and keep everyone locked in.[17]
BEYOND THE LADDER
I want to end by returning to the frayed edges of the composite. The poor chroma of StreamArt is not yet an exit from the platform condition, but nor is it, as UKRAiNATV insist, a celebration of glitch for its own sake. I see it as an encounter between the corporate stack and other values, other ways of relating. If an early promise of video was instant feedback, the platform has transformed this potential into an administered, profit-seeking condition. Yet StreamArt follows the lineages of experimental video, exploring the transformation of that spirit inside today’s streaming vortex. Something of the experiment survives – the ladder sees to it – but that doesn’t tell us what survival is for. For that, we might need to consider a migration to other horizons, away from the stack towards more autonomous infrastructure. But then, what would StreamArt need to sacrifice to make that trip, and would it survive?
—
[1] CDI-TV is a collective comprising Carolina Bandinelli, Michael Dieter and Craig Gent, based at the Centre for Digital Inquiry (CDI), University of Warwick. See https://warwick.ac.uk/fac/cross_fac/cdi/
[2] Together and apart is a formulation from Branden Hookway, Interface, Cambridge, MA: MIT Press, 2014.
[3] Sean Cubitt, Videography: Video Media as Art and Culture, London: Macmillan, 1993, pp. xv.
[4] On instantaneity as a recurrent, but non-essential feature of video, see Dieter Daniels and Jan Thoben’s monumental Video Theories anthology, which gathers these authors and offers a typology of instantaneity: Live Remote Image, Live Self-Image, Live Feedback, Instant Replay and Instant Replay Self-Image. They also note that the instantaneous video image “can only be described in relations and differences: it is specific to the medium but does not offer a general, essentialist definition of it.” Dieter Daniels and Jan Thoben (eds) Video Theories: A Transdisciplinary Reader, New York: Bloomsbury Academic, 2022, p. xxvii.
[5] Paul Ryan, ‘Videotape: Thinking about a Medium,’ in Daniels and Thoben (eds) Video Theories, pp. 28-31.
[6] Maurizio Lazzarato, Videophilosophy: The Perception of Time in Post-Fordism, trans. Jay Hetrick, New York: Columbia University Press, 2019, p.1.
[7] Maurizio Lazzarato, Signs and Machines: Capitalism and the Production of Subjectivity, trans. Joshua David Jordan, Los Angeles: Semiotext(e) Foreign Agents Series, 2014.
[8] Michel Serres, Hermès IV: La distribution, Paris: Minuit, 1977.
[9] Konrad Wojnowski, ‘UKRAiNATV – Streaming Philosophy,’ Tactical Media Room, Institute of Network Cultures, 26 April 2024; https://networkcultures.org/tactical-media-room/2024/04/26/ukrainatv-streaming-philosophy-by-konrad-wojnowski/
[10] Geert Lovink, Platform Brutality: Closing down Internet Toxicity. Amsterdam: Valiz, 2025.
[11] Tiziana Terranova, ‘Technoliberalism and the Network Social,’ Theory, Culture & Society 41.7-8 (2024): 89-104.
[12] Expanded Publishing: https://expandedpublishing.net/
[13] Carolina Bandinelli, Michael Dieter and Craig Gent (eds) Hybrid Encounters: CDI-TV Season One. Coventry: University of Warwick, Centre for Digital Inquiry, 2026; https://warwick.ac.uk/fac/cross_fac/cdi/cdi-tv_hybrid_encounters.pdf
[14] Hito Steyerl, ‘In Defense of the Poor Image,’ e-flux Journal 10 (2009), https://www.e-flux.com/journal/10/61362/in-defense-of-the-poor-image/
[15] Ned Rossiter, ‘Logistical Worlds,’ Cultural Studies Review 20.1 (2014): 53-76
[16] Harun Farocki, ‘Phantom Images,’ trans. by Brian Poole, Public 29 (2004): 12-24; https://monoskop.org/images/f/f8/Farocki_Harun_2004_Phantom_Images.pdf. See also Jussi Parikka’s contemporary extension, Operational Images: From the Visual to the Invisual. Minneapolis: University of Minnesota Press, 2023.
[17] For a critique of this political epistemology, see Orit Halpern and Robert Mitchell, The Smartness Mandate, Cambridge, MA: MIT Press, 2022.