(expanded version of talk given at the opening evening of INC Exit Fest at UvA Spui25 in Amsterdam, June 22, 2026)
Tonight, I would like to cast the idea of flat world theory to you and introduce the concept of the logistical episteme. Flat world theory – this is not to be confused with the Latourian world of flat ontologies where conflictual confrontations of the political evaporate into a chain of empty signifiers and relational assemblages. But we might be talking about what Anna Kornbluh in her critique of the immediacy of style in too-late capitalism calls “Knausgaard’s acid-brilliant flat ontology: everything is everything.”[1] To be sure, everything is everything is just another variation of flat ontologies. It’s the acid-brilliance that unsettles the order of things.
Flat world theory takes the world for what it is in the logistical age of AI: flat as a mat.
Screenshot from the ChatFOS research experiment by Zoe Horn, Anna Munster, and Liam Magee, 2023.
As much as we might find purchase in the idea of performative actions constituting the propensity of things, such a precept disavows the media question, which we can extend to the infrastructural question: namely, that propensities may in fact precede actions. This is not to say there’s an underlying essence, say, to something like ‘the internet’, or radio, or newspaper for that matter. But it is to acknowledge that materialities of communication have signalling capacities inherent to both their design and the matter of their composition. And it’s also to acknowledge that there are histories – in ideas about how to do things and in technological forms and artefacts that have, for instance, legacy architectures that infuse and carry over from one update to the next.
History precedes and is also baked into the scene of encounter where things perform actions and build relations. History is anathema to the theatre of flat ontologies.
Melbourne logistics zone. Photograph by Ned Rossiter, 2025.
++
Certainly, there is strong scientific truth and conceptual merit in the notion of the Earth as a system. But do we call it media? At a smaller scale, are elemental media (clouds, fire, wind, sun, oceans) really media? For sure, we recognize and experience the ways in which these components of planetary life have mediating capacities. But is this the same as saying they are media? Do they communicate in ways that engineer our cognitive capacities, as Bernard Stiegler was at pains to argue?
Stiegler, by the way, was not the first to make this observation – we find a long tradition in France that takes an interest in the relation between technology and material conditions, specifically as they pertain to cognitive processes. The early- to mid-twentieth century anthropologist André Leroi-Gourhan was particularly influential for Stiegler’s thinking. Stiegler notes that “Leroi-Gourhan in fact says that it is the tool, that is, tekhnē, that invents the human, not the human that invents the technical. … the human invents himself in the technical by inventing the tool – by becoming exteriorized techno-logically” in the form of tools.[2] Further, “‘externalization’ is the continuation of life by means other than life.”[3]
We can also call that tool infrastructure.
Let’s turn this idea around to help clarify what we mean by the media question. Media, indeed, are another tool or technical device external to the human, even when organically integrated into our very bodies as chips, implants, and the like. Our bodies and brains have little control over these devices. Power of this kind resides outside the body and is typically concentrated in what the Canadian communications theorist Harold A. Innis termed monopolies of knowledge.[4] In other words, the media question also requires attention to the political economy of technology.
Digital Free Trade Zone, Malaysia. Photograph by Ned Rossiter, 2025.
This matters for how we consider the politics – indeed the geopolitics and also geoeconomics – of critical infrastructures and their organization of society and culture.
++
It almost goes without saying that in analysing the ever-shifting ground of the present we are in a kind of constant diagnostic battle to identify substantive tendencies in the face of unrelenting contingencies. For all the market hype around AI and associated infrastructures, massive builds in data centres chief among them, an event as destructive as the US-Israel war on Iran makes clear how fragile any such certainties may previously have seemed. Infrastructure might be concrete until it is blasted away by a ballistic missile (or a drone).
Jurong West, Singapore. Photograph by Ned Rossiter, 2017.
These days we discover it’s not only bombs that turn bunkers to dust. We read that the feasibility of data centres is seriously in question across the Gulf states, as well as elsewhere in the world, due to the impact of energy cuts.[5] Sure, electricity grids can be powered by coal, nuclear, and increasingly more sustainable energy sources. But data centres are backed up with diesel generators and many power stations require gas turbines to generate electricity. And power plants and electricity grid networks are managed by computational systems that often incorporate AI systems to forecast and optimize energy load, distribution, and demand in near real-time.
War, then, is a central part of the current infrastructural conjuncture. Think, for example, of helium. Whoever heard of it before? Now we all know how critical it is to the manufacturing of wafers for microchips. And we know that around 38.8 per cent percent of the world’s helium comes out of Qatar – a key ingredient for the coolant derived from waste liquid natural gas (LNG) plants, which is then used in the etching of wafers during the manufacture of semiconductors in Taiwan, Japan, Singapore, and South Korea. As we know well, these chips are then used in many commercial applications and digital technologies such as AI, smartphones, EVs, MRI scanners, data centres, and the production of fibre optics and satellites in telecommunications and aerospace industries.
Source: X
And we know that Trump’s failed blitzkrieg of Iran has resulted in massive blockages in shipping through the Strait of Hormuz, wreaking havoc on livelihoods and economies in pretty much every country you can think of. And we know that all those AI data centres piling up around the world are in jeopardy – not only because they are blowing bubbles that are gonnna burst, but much more directly they are at huge risk of not being built at all precisely because of the nexus between war, infrastructure, resources, and the economy. The horrendous annihilation of civilian life and environmental devastation are considered merely as collateral damage on the sidelines of mission control within the logistical episteme.
Whether it is the tropics of AI that we might associate with data centres and chip manufacturing and packaging in Singapore, Malaysia, and Indonesia’s Batam Island, or investment by Chinese firms such as Huawei and ByteDance in Brazil’s data centre and AI sectors, from a logistical perspective these AI climate zones are inseparable from the Middle East’s deserts of war.
++
So what, then, is flat world theory? It’s something that emerges from the media question as it pertains to computational interoperability. The fantasy of integrated hardware and software systems that comprise vertical stacks all the way down. It’s the promulgation of a conviction that computerization is equivalent to networking rather than notworking (Lovink). Flat worlds are found across many of the organizations, companies, and government agencies we might work in. They all bought into the seduction trap of Silicon Valley. Install the system and efficiencies and organizational optimization will follow. Total bullshit!
Electronic waste site, Hong Kong. Photograph by Ned Rossiter, 2017.
Flat world theory is the mirror of control. We can use it to call out the crap. Let’s produce the memes and flood the feeds with next-level nonsense! Here, we can find inspiration in Gabriele de Seta’s wild and nutty work on Italian Brainrot. Along with truly stomach churning digital culture he digs up in kooky corners of TikTok insanity, Gabriele points us to something critical for our understanding of the geopolitics and geoeconomics of production networks and related infrastructures.
Indeed, there’s an entire field in economic geography that investigates global production networks and material political economy.[6] A chief interest in this literature is to identify spatial dependencies across sites of production, distribution, and supply necessary for the manufacturing of microchips, for instance. Here we can find out about the international division of labour as it relates to the extraction of rare earths elements. Palladium, cerium, neodymium, and holmium are among the thirty or so critical minerals required by AI and digital technologies. We learn also about the factories across South East Asia where low-end manufacturing and packaging of microchips happens as part of industrial-organizational geographies usually centred around East Asia’s top foundry fabs of Samsung and TSMC in South Korea and Taiwan.
Gabriele de Seta shows us how something as trashy and freewheeling as Italian Brainrot also moves across global production networks with animated AI replicants getting churned out in a frantic fashion in dedicated AI data centres then finding material form as junky tourist kitsch in $2 shops in Mongkok. Gabriele also happens across more Italian Brainrot, this time in Oslo from a toy dispensing machine. Noting the many labels stitched into the figure, he was able to trace its journey across manufacturing sites and global supply chains, having been knocked out of a factory in Guandong, filtered through a toy company based in Dubai before importation in Denmark. Much weirder and perverse synthetic characters are mushroomed out by social media users and AI slop factories in China. Italian Brainrot tells us the remarkable story of distribution and transmutation unhinged from proprietary forms of control.
My point in bringing in the Italian Brainrot story is that it tells us that whether it’s the high-end billions of dollars market of chip manufacturing or the down-at-the heels world of AI kitsch-culture, similar networks of production and supply are at work. It’s all about infrastructure and logistics, regardless of price point.
Digital Free Trade Zone, Malaysia. Photograph by Ned Rossiter, 2025.
++
As the complexities of AI systems multiply across notionally distinct domains of governance and markets, engineering and law, we start to get a sense of an infrastructural drift toward what Berlin-based artist and writer Hito Steyerl calls “the political economy of entropy” – the point at which order and organization tip over into noise, randomness, disorder, instability, and disruption.[7] For Steyerl these kinds of entropic conditions push societies toward “political fragmentation … targeted identification, micro-labelling and desolidarisation.” Combined, such forms of “entropic engineering,” writes Steyerl, make “workers in digital industries” assume the role of “noise particles,” “burnt out,” and “vulnerable to exploitation.” These are the sorts of machinic conditions Steyerl associates with the political economy of entropy in the age of AI.
She finds inspiration in Bernard Stiegler’s concept of negentropy, a hope or belief or political commitment that strives to “create order by way of organisation,” that considers “denoising” as a technique of protests and strikes that “are not just disruptions of ordinary routines and procedures but effect the disruption of disruption.” In other words, a form of what Steyerl also calls “reverse entropy” or “organisational negentropy” in which the protagonist “appears in absence and negation: as a hole in the fabric of entropic engineering.”
By my reading, the idea of negentropy is imbued with a certain nostalgia for a era that never existed. As an appeal for a return to some kind of balanced existence in a world well and truly out of kilter, such yearning is no real surprise. But might it not be another variant of flat world theory? To counter the disorder of entropy with the ordering propensity of negentropy is something of a zero sum game, an assertion of equivalence where there is none. Yet there is something in Steyerl’s proposition of reverse entropy in as much as it constitutes a political intervention aimed at puncturing the flat world of entropic engineering.
Digital Free Trade Zone, Malaysia. Photograph by Ned Rossiter, 2025.
What I’ve tried to sketch here is a parable for INC 2.0. As an autonomous organization at the cross-roads, INC needs to make a decision about the path ahead. How does INC excavate the future without also taking into account the critical question of designing infrastructures of possibility? These are not flat worlds, but exactly the opposite: a hostile refusal of and attack against the power of flatness.
Might we find in the midst of this disruption of disruption a radical exit from the logistical episteme and its insistence on the optimization of performance and the ordering of disorder?
Beijing. Photograph by Ned Rossiter, 2026.
[1] Anna Kornbluh, Immediacy or, The Style of Too Late Capitalism (London and New York: Verso, 2023), 151.
[2] Bernard Stiegler, Technics and Time, 1: The Fault of Epimetheus, trans. Richard Beardsworth and George Collins (Stanford, Cal.: Stanford University Press, 1998), 141.
[3] Bernard Stiegler, “Leroi-Gourhan: The Organised Inorganic,” Cahiers de Mediologie, no. 6, 2e Semestre (1996).
[4] Harold A. Innis, The Bias of Communication (Toronto: University of Toronto Press, 1951).
[5] Indranil Ghosh, “Big Tech’s Gulf megaprojects are trapped between two war choke points,” Rest of the World, 4 March, 2026, https://restofworld.org/2026/us-iran-war-gulf-ai-submarine-cables/.
[6] See Henry Wai-Chung Yeung, Interconnected Worlds: Global Electronics and Production Networks in East Asia (Stanford, Cal.: Stanford University Press, 2022).
[7] Hito Steyerl, Medium Hot: Images in the Age of Heat (London and New York: Verso, 2025).