My research examines the history and politics of technological modernity. I am specifically interested in the ways the material infrastructures of the modern world embody and uphold relations of power and domination, despite promising to liberate humanity from scarcity and toil.

My doctoral research centered on the longue durée of industrial modernity and what the early Frankfurt School theorists called “instrumental reason,” tracing its roots to the nineteenth century. My dissertation, Karl Marx, Philosopher of Technics: Critique, Alienation, and the Materiality of Social Power, grounded this inquiry in mid-19th-century debates on technology and political economy that decisively shaped the first appraisals of industrialization, including Marx’s central intervention. Drawing on understudied materials from the Marx-Engels Gesamtausgabe (MEGA²), I reconstructed how Marx came to understand the relationship between capitalism and the modern technical system, against the backdrop of a techno-determinist common sense growing out of the Industrial Revolution. In particular, my study followed the historical-critical edition of Marx’s technological notebooks and the associated MEGA² scholarship to analyze the formative influence of Charles Babbage and Andrew Ure in the composition of Part IV of Das Kapital, dedicated to the capitalist transformation of the labor process. On this basis, I argued that, contrary to prevailing economicist and productivist readings, Marx theorized industrial development and its constitutive mechanism – the “subsumption” of the labor process by capital – as an alienated organization of human-technical relations, whereby capitalist power becomes embedded in the very conditions of human life.

Machine Hall, The Great Exhibition of 1851

Karl Marx (1845), Notes and excerpts from the French translation of Charles Babbage’s On the Economy of Machinery and Manufactures (1832)

As I revise this work into a monograph, my ongoing research traces how the core dynamic of subsumption has steered the course of the information age and is reasserting itself in the contemporary AI project – particularly in its claim to instantiate intelligence as such through the statistical modeling of human language use. Against the market-driven presentism and techno-determinist ideology that pervades the current debate, I seek to understand machine intelligence in its longue durée, as the latest development in the long history of industrial automation that undergirds late capitalism’s digital turn. My wager is that the computational infrastructure that has come to mediate so much of contemporary life is best understood not as a “means of production,” in the classical Marxist sense, but as a subsumptive technique that generalizes the industrial form of the labor process – that is, the organization of human activity and social cooperation according to a logic of labor-time abstraction.

Since the postwar period, the advent of the modern electronic computer has been celebrated for reintroducing human feedback into the unthinking, mechanical apparatuses of industrial modernity. By making external input and self-regulation integral to machine operations, the computer promised to reverse the decline of reflexivity that 20th-century critics – from phenomenology and existentialism to early critical theory – ascribed to mechanization and the reign of instrumental reason. Yet this optimistic self-image of the digital age has consistently faced scrutiny within media and technology studies – most pointedly in light of Gilles Deleuze’s late concept of “control societies,” which directly linked the rise of information machines to a new diagram of power. Building on my previous inquiry into the relationship between social and technical forms, my current project takes up, more specifically, the relationship between computation and the industrial labor process, turning to the 20th-century history of computing, postwar appraisals of machine intelligence in the human sciences, and their ideological afterlives. Paying particular attention to the development of early real-time, “command-and-control” computer systems, such as the Whirlwind computer and SAGE, and their connections to mid-century debates within the cybernetics movement and the nascent field of AI, I aim to ask how the social ontology of computational technologies is bound up with capitalist forms of cooperation, and whether the socialized intelligence they harness makes possible alternative, non-alienated social relations.

Whirlwind Computer Room Floor Plan, Barta Building, MIT (1951)

Components of the IBM Semi-Automatic Ground Environment (SAGE) computer system (1951-1963)

Work in Progress: Computation and Social Form

What is a computational system and what form of human grouping begets its existence? Or put another way, what is the form in which modern humans confront materiality such that it requires an automated system of “control and communication” (Wiener)?

A computer is, according to most modern definitions, a system that performs sequences of mathematical or logical operations on inputs, according to rules specified in advance. So defined, many technical systems preceding the digital electronic computer would qualify as “computational,” from the Jacquard loom to the group of human calculators employed in the production of logarithmic tables until the late nineteenth century. Yet it is only under conditions of advanced industrial development that computational systems have come to organize social life as a whole, embedding information-processing operations in nearly every facet of the modern lifeworld. From scientific research, manufacturing, and the circulation of goods and people to the tracking and management of consumer desires, behaviors, and political sentiments, there seems to be no aspect of life in contemporary society that is not either directly or indirectly mediated by computational processes. It is, therefore, unsurprising that the digital has come to function as a technological synecdoche for late capitalism itself – as a new form of “enclosure” that, like its analog predecessors, abstracts and elides the heterogeneities of the material world to render it exploitable.

Yet is control and domination by logico-mathematical principles all there is to computational processes and the machinic modes of intelligence they enact? While the digital computer has emerged from and ultimately serves, in its present form, capitalist domination, the notion that computational technologies are inherently functional to that purpose deserves further investigation. As contemporary media-theoretical accounts of computation propose (Hansen, Parisi, Hui, Fazi), computation understood as a general material-cosmological operation is not reducible to the computer-regulated networks of communication and control it has become narrowly associated with in our times. The way capitalism deploys computational processes does not, in other words, exhaust the world-disclosing potentials that digital technologies exhibit, particularly their capacity to widen the scope of human sensibility and self-understanding beyond received anthropocentric assumptions and norms.

In my current research, I aim to work through this tension by studying what I am calling the “social ontology of computation.” I hypothesize that only by grasping the historically specific forms of social cooperation that computational technologies mediate is it possible to adjudicate between their different valences. As I discuss next, this entails clarifying the relationship between abstract forms of social mediation and alienation – a problem for which we find indispensable resources in the work of thinkers like Marx and Simondon, who recognized it as the central pathology of industrial society. The question I wish to pose is, ultimately, whether abstract social mediation is itself the source of alienation, or whether alienation arises from the social form through which abstraction is organized.

Abstraction, Alienation, and Social Ontology

I understand the social ontology of a technical system to concern the human processes at work in its genesis and evolution (without this implying that the technical is reducible to human intention and use).

Broadly conceived, social ontology examines the modes of being of the social – that is, of the “transindividual” reality that is immanent in individuals and their collective activity, and that nonetheless exceeds and conditions their determinate existence. As a mediator of this reciprocal relation between the individual and the collective, technology does not merely instantiate preexisting human ends, but is implicated in the process of constituting the human as such. Conversely, since no medium is independent of the processes it mediates, human life and its historically changing shapes provide the conditions through which technology’s immanent tendencies are realized.

My working hypothesis concerning computation is as follows: contemporary computational systems are formally determined by the industrial labor process, understood as a historically distinct arrangement of the interface between humans and machines arising with capitalist modernity. More precisely, the industrial labor process provides the social form within which the technical features of computation become intelligible and materially realizable – without, for this reason, mechanically causing the modern computer to exist or determining its empirical content.

By “industrial,” I do not mean the specific technologies and industries of a bygone Victorian or Fordist era, now seemingly superseded by a de-industrializing world, but a distinct organization of human labor and social cooperation modeled on commodity relations and the reproduction of capital. What distinguishes this industrial form is not, as the common understanding would have it, the productive employment of humans, machines, and nature for maximum physical output, but rather the allocation of abstract labor-time in the most value-efficient manner. Importantly, this should not be confused with the “profit motive” to which orthodox Marxism accorded explanatory primacy; nor does it imply that all social activity directly assumes this form, or that the concept of “industry” narrowly refers to the factory and the production of physical goods. State bureaucracies, the military-industrial complex, the culture industry, and the array of modern institutions that Michel Foucault characterized as “disciplinary” equally belong within this historical arrangement, whereby human bodies are organized and rendered productive against a common standard of efficiency.

As a growing body of scholarship in the history of computing has shown (Daston, Schaffer, Pasquinelli), mechanical calculation developed in close relation to industrially organized mental labor, suggesting a common operational logic. Indeed, if the industrial coupling of humans and machines necessitates a systematic regimentation of work and value, then computation is the mechanism by which the modern technical system makes this requirement operational at scale. What is more, as informational processes become more granular, recursive, and operative across the whole system, they also become increasingly inscribed in the self-regulation of its parts. This technical tendency toward machine autonomy is not, contrary to conventional Marxist and Schumpeterian accounts, sufficiently explained by extrinsic economic imperatives – such as profitability, labor-saving, or competitive pressure, which are compatible with other technical paths. Instead, as labor-process theorists have shown (Noble, Braverman, Panzieri), it results from the mechanism through which the industrial labor process formally determines the material content it subsumes: the operationalization of the division of labor.

In its drive to allocate abstract labor-time with maximal precision and efficiency, large-scale industry models every process as a sequence of rule-governed operations that can be decomposed and formalized independently of physical embodiment. To subsume an activity under the industrial form is, in this regard, to sequence its mechanical and human contributions appropriately yet without making ontological distinctions between them, as functional equivalents capable of rule-governed conduct. In fact, labor and rule-following become coextensive in this context, and machinic intelligence thus comes to rest on an algorithmic automatism, formally encapsulated by the universal Turing machine. Despite the contemporary turn to probabilistic reasoning and the unprecedented possibilities that machine inference has opened up, the “synthesis of word and number” (Weatherby) achieved by large language models does not fundamentally depart from this tendency and its underlying logic of labor-time abstraction. Machine autonomy, as the broader history of automation attests, does not constitute a rupture but an immanent development of the industrial labor process – even as machine learning algorithms now automate programmability itself, the “rule-writing” process that previous waves of automation entrusted to the human engineer, recomposing it as data annotation and evaluation.

The organizing questions of my ongoing research are as follows. If contemporary computational systems are the “transindividual” medium of social cooperation based on commodity relations, do all similar modes of abstract mediation entail reification, as a certain Lukácsian line of thinking (Adorno, Sohn-Rethel, Postone) would have it? In other words, does the social ontology of computation necessitate oppressive forms of valuation like the capital relation? Or is what critical theorists call reification no more than the functional differentiation of autonomous subsystems, inherent in all social formations, as systems theory claims? Between Lukács and Luhmann, a third Simondonian path might suggest that alienation is not a function of mediation itself but of a truncated collective individuation. Indeed, an alternative system of social cooperation based on universal need satisfaction would still seem to require its own means of commensuration for comparing and reconciling non-identical contents, computation thereby grounding human flourishing rather than a self-moving abstraction. Whether these questions can be answered conclusively is what I am presently contending with, but I hypothesize that the unrealized tendencies that computational systems exhibit, even when enfolded in an oppressive social world, suggest a non-alienated form of abstraction implicit in general social cooperation.