10. Carbon and silicon: contribution to a critique of political economy

Daniel Ross

What is sometimes called the “hyperindustrial” economy – defined by digital technology and disruptive marketing – is based on those two atomic elements (in the chemical sense of the term) that are carbon and silicon. They are also the two elements of hyperindustrial society in another sense: in the sense involved when we refer to water as the element of the fish – and where today these seem to have become toxic for those who try to live within this elemental milieu.

Carbon becomes the center of the industrial economy with the invention of “heat engines” (as Sadi Carnot called them), while silicon becomes the element of control technologies and the exploitation of individual and collective memory. To what geo-economio-politics of hypercontrol does this lead, especially in the light of China’s rise to the status of hyperpower? And in what way do these questions necessitate the perspective of a general theory of thermodynamic, biological and informational entropy that aims to rethink the conditions of decision-making in the Anthropocene?


We are confronted in the twenty-first century with an array of serious problems but among them two immense challenges stand out: on the one hand, those problems presented by carbon technologies, and, on the other hand, those posed by silicon technologies. While it may seem that nothing can trump the planetary threat of climate change, in fact both of these challenges involve existential threats and dangers amounting to what is sometimes called ‘extinction risk’, not least because these two challenges are absolutely inextricable. We believe that there is a widespread intellectual and political blind-spot about the economic, political, psychological and sociological significance of the vast technological transformation that has unfurled across the past quarter of a century. More specifically, it is today crucial to understand the complex and fundamental ways in which seemingly irreconcilable questions of economic and ecological sustainability relate to and are compounded by the transformation of computation, information, network and algorithmic technologies.

Carbon technologies

Hominims acquired the ability to create and use fire as early as the Lower Palaeolithic and the controlled use of carbon combustion became common in the Middle Palaeolithic. From that moment, the beings that would become ourselves found themselves within a fiery element defined by the capacity for artificial, controllable energy production and consumption founded on the flammability of organic materials. From the moment cooking was invented, this capacity was a matter of the potential to produce and consume energy in order to do work, thereby opening the possibility of ‘ways of life’, or what Marx called a ‘mode of consumption’: ‘the hunger that is satisfied by cooked meat eaten with knife and fork differs from hunger that devours raw meat’1. Both dangerous and beneficial, controllable within the risks of being extinguished or turning wild, this first symbol of technics was also the first object of care, long before the Neolithic Revolution.

In addition to warmth and cooking, the development of the controlled use of carbon combustion gave rise to other technologies, such as smelting, forging and gunpowder. But the modern history of carbon technologies obviously begins with the invention of heat engines powered by hydrocarbons derived from fossilized organic matter. More specifically, it begins with the external combustion engine, and more specifically still with the industrial (or thermodynamic) revolution that was set off by the steam engine envisaged by James Watt, which he patented in 1781 and which was to transform manufacturing and rail and maritime transport throughout the nineteenth century. In the twentieth century, fossil fuel power plants linked to electricity grids would further vastly transform both production and consumption, and automobiles equipped with internal combustion engines would transform road transport and make possible the rise of global aviation. The combustion of hydrocarbons, however, inevitably releases a significant level of ‘metabolic products’: while for the ten thousand years prior to the industrial revolution the global atmospheric CO2 concentration was 280 parts per million, in 2018 it currently stands at 410 ppm, with annual emissions and concentrations continuing to increase2.

Silicon technologies

The first integrated circuit was produced in 1958, the first CPU in 1971, the Apple II and Commodore PET home computers entered the market in 1977, the Microsoft Windows ‘operating environment’ was first released in 1985, the World Wide Web was opened to the general public in April 1993, Amazon was founded in July 1994, the domain name google.com was registered in September 1997, the Tencent and Alibaba conglomerates were founded in 1998 and 1999 respectively, the Facebook social network was made universally open in September 2006 (with active users rising from 100 million in August 2008 to two billion in June 2017), the capacitive multi-touch smartphone known as the iPhone was launched in June 2007 and Uber’s mobile app and transport services were officially launched in July 2009. It is notable that this timeline of significant dates increasingly focuses on consumer-based silicon technologies, reflecting the vast entrance of these transformational technologies into the consumer market over the past forty years. It is also notable that we have chosen to end it in 2009, reflecting that the last decade has seen a period of consolidation and monopolization of the silicon economy in the hands of a small number of super-giant corporate players.

Today, it has become transparently clear to everyone that silicon technologies have transformed every aspect of production and consumption, along with scientific and technological research of every kind, and this is especially so in the quarter of a century that has transpired since the internet became public and global. All of this amounts to a vast ‘disruption’ of the technical system, along with every other psychosocial and institutional system.

Retentional technologies and the industrial capitalism of production

If the elemental function of carbon technologies fundamentally consists in the production of chemical energy in order that it can then be transformed into mechanical or electrical energy and consumed as work, then the elemental function of silicon technologies fundamentally consists in the production of an artificial memory that, too, can be put to work in manifold ways. Silicon technologies are retentional technologies (to borrow a term from Husserlian phenomenology). In Stiegler’s work, this very long history of retentional technologies (or what he prefers to call hypomnesic technologies) has been explored in detail and with respect to a wide variety of dimensions. If we here prefer to refer to silicon technologies – while keeping in mind the mnemotechnical history that extends through cave painting, the invention of writing systems (including alphabetization, which remains an almost unchanged standard from the Roman Empire to the Digital Leviathan), the printing press, the phonograph, the radio, cinema, analogue television and the becoming-digital of everything that we now see unfolding with silicon technologies strictly speaking – if we refer to silicon technologies, therefore, it is only because the proliferation of uses, services and functions associated with this latest stage of memory technology seems so greatly to exceed the mere ability to ‘record the past’. And yet, this is precisely the basis of all of them.

The industrial revolution whose possibility we previously ascribed to Watt’s steam engine could never have occurred without retentional technologies of a kind we have hitherto failed to mention: those technologies by which the continuous gestures of workers of all kinds were broken down analytically into their discrete elements, in order to be then programmed back into machines powered by the heat engines of Watt and his successors: the paradigmatic case of such a machine is Jacquard’s loom, but a thousand examples could no doubt be cited. The basis of this analytical process is what Stiegler refers to as ‘grammatization’, the process of turning something temporal (like speech) into something spatial (like writing), by turning the continuous into the discrete, on the basis of which it can be analysed and reproduced.

The noetic, political and economic consequences of grammatization can be to support new forms of knowledge, but it can also lead to what Stiegler calls ‘proletarianization’ (drawing on Gilbert Simondon’s reading of the Grundrisse’s ‘fragment on machines’). If proletarianization has in traditional Marxist discourse been understood to refer to the systematic separation of workers from the means of production, Stiegler’s use of the term draws attention to the way in which those means firstly consist in the knowledge possessed by workers and transmitted inter-generationally. It is this knowledge that is literally removed from the minds of weavers and programmed into Jacquard’s loom. In addition to the ownership of the energy-production capacities of the heat engine, what in fact made the rapid acceleration of the industrial revolution possible was thus the ability of the capitalist to dispossess the worker of the knowledge of how to make things, knowledge that was then turned into information and recorded and exploited in the retentional technologies of machines: it is here that the history of industrial automation and artificial intelligence truly begins.

Industrial capitalism based on production thus arises from the concentration of carbon technologies in the hands of capital, but equally from the capitalist acquisition of retentional technologies through which workers, systematically dispossessed of knowledge, become labourers, that is, servants of the machine. From this vast process is born that great division between capital and proletarianized labour on the basis of which Marx and Engels would construct a revolutionary politics. In fact, of course, this founding moment of the industrial revolution was only the first step of a history that would continue through many chapters, including ones that Marx could never have anticipated: one key way in which to understand this set of chapters is as the unfolding of the epochs of grammatization.

Protentional technologies and the hyper-industrial capitalism of consumption

This is precisely the realization that came to capitalist producers at the beginning of the twentieth century. For Marx, the spread of machines (powered by carbon technologies and programmed by retentional technologies of grammatization) amongst the capitalist class was bound to make it increasingly difficult for any one capitalist to maintain an edge over others, leading to his diagnosis of a tendency of the rate of profit to fall. Economists ever since have disparaged this analysis, above all on the grounds that it is not what is observed in the economic history that has unfolded since it was described by Marx, ‘natural’ boom-and-bust ‘cycles’ notwithstanding. Indeed, this history does not seem to confirm Marx’s analysis. But this may be the result less of analytical error than of a fundamental transformation of capitalism resulting from this tendency, even if the solution to this problem is itself only a postponement of this tendency.

The essence of this ‘solution’ was the realization that it is possible to create new markets, not just by geographical expansion, but through the possibility of manipulating consumer desire and therefore consumer behaviour. If capitalism is a perpetual economic competition giving rise to perpetual technoscientific innovation, this is not just a matter of R&D and production: it is also a matter of the socialization of that innovation – all those processes through which new products are taken up by consumers, by which they are adopted. The shift to a hyper-industrial capitalism of consumption was in part a matter of the new organization of consumption, but the large-scale investment required to achieve the productivity gains to be realised from mass production was feasible only if consumer behaviour could be more or less reliably predicted, which is to say, produced: for this new consumer market in transport vehicles powered by internal combustion engines to succeed, it was necessary to invent public relations, or in other words, marketing.

As Stiegler has shown on many occasions, this invention was made possible not just by the discovery of this ‘idea’, but by the development of new forms of grammatization, and specifically the ‘grammatization of the sensible’ inaugurated with audiovisual technologies such as radio, cinema and television. It is not technological change as such that Marx could not anticipate, but the significance of the new analytical and programming possibilities opened up by these new retentional technologies. With these powerful new tools that could be used to access and influence the minds of potential consumers on an industrial scale, it became possible to completely transform the basis of profit-making in industrial capitalism, by constantly manufacturing the market for the new products that could then be constantly introduced and updated.

By accessing consciousness and targeting the unconscious, marketing and its associated technologies and techniques have progressively learned how to make consumer behaviour controllable, by interpolating (in the literary sense) tertiary retentions into the stream of consciousness. The basis on which it can do so, however, depends on reducing desire as much as possible to a calculable phenomenon, which is to say, grammatizing the relationship to the future. In other words, this amounts to a grammatization of protention, Husserl’s term for my immediate expectation, but expanded here to include every form of motive, reason, expectation, dream and desire.

This in turn involves a detachment of desire from everything incalculable, incomparable and long-term (including every form of education and inter-generational transmission), inducing a regressive tendency that aims instead only at the finite and short-term goals of the consumer behaviour required by the market. But this ultimately risks being self-destructive for the consumerist model itself, setting up a tendency for the libidinal economy (on which the macroeconomic ‘perpetual growth model’ fundamentally depends) to collapse, as libidinal energy is depleted: the ability to stimulate the perpetual increase in consumption required by the consumerist economy is thereby threatened. It is ultimately for this reason (along with the aporia of sustainability) that consumerist capitalism can be but a postponement of Marx’s diagnosis with regard to the rate of profit.

Silicon technologies and the ultra-industrial capitalism of algorithmic platforms

The protentional grammatization technologies of the twentieth century had only limited means of accessing the information and data that is necessary in order to calculate and predict the relationship between, firstly, grammatized content (for example, a television commercial that, in Husserl’s terms, amounts to a kind of industrial temporal object), secondly, protentional conditioning, and thirdly, consumer behaviour: the clearest indicator was ultimately the success or otherwise of a marketing campaign. But with the introduction of silicon technologies that now dominate the twenty-first century, this question is fundamentally transformed, because the consumers of such grammatized content are ceaselessly and immediately sending data back to producers. On the basis of such data, producers can ever more finely calculate the relationship between particular content and particular responses from particular ‘kinds’ of users. The extreme speeds at which these processes occur in contemporary algorithmic silicon systems means that it is also possible for these producers to adjust content in a very rapid and targeted way that was simply impossible in the twentieth century. This speed exceeds that of noetic processes themselves, and this rapid exchange and algorithmic control of vast amounts of user data gives rise to a kind of informational and protentional shock wave, analogous at the noetic level to the ‘sonic boom’ produced at flight speeds above Mach 1.

Every major consumer platform today utilizes powerful algorithmic techniques of this kind in order to absolutely maximize their ability to performatively influence consumer behaviour. Furthermore, global ‘platforms’ such as Alphabet and Facebook are now among the largest corporations on the planet and have become so through the new market they have created for the vast amounts of data generated by their users. If the capitalism of analogue audiovisual technologies was already hyper-industrial and performative (in Austin’s sense), then the new market of platform capitalism based of silicon technologies, user profiling and social networking is highly performative and can thus be considered an ultra-industrial capitalism of algorithmic platforms3. But this only intensifies the deleterious effects of such processes on the libidinal economy of consumers. And this in turn is bound to intensify the self-destructive tendencies of the consumerist macroeconomy, since it ruins the very basis of its ‘success’: the control of desire.

The anti-politics of ultra-industrial populism in the Entropocene

Behind such a paradoxical intention to produce consumers lies the even more paradoxical belief that this mass of consumers can continuously drive the engine of the global economy like a perpetual motion machine, and drive it to new heights. But perpetual motion is a myth based on the notion of an abstract machine that is thermodynamically impossible, and the ‘heights’ to be reached are in this case transparently at odds with the unambiguous imperatives declared by the IPCC. But in addition to that, the billions upon billions of bytes of data gathered from consumers by producers and platforms, fed into increasingly powerful and increasingly intelligent automated algorithms designed to calculate and control behaviour according to the imperatives not of the IPCC but solely of the market, has an extremely ruinous effect on the psyches of the individual consumers of whom this market is composed (who are today targeted almost from birth, if not from before birth), giving rise as it does to an infernal spiral of consumerist addiction.

Evidence abounds throughout the industrialized democracies of the political consequences towards which this ruination tends. And these consequences are intensified by the fact that all these performative techniques are applied also in the political realm. If, as Stiegler suggests, this entails the replacement of the adoptive performativity of ‘democracy’ with the adaptive performativity of ‘telecracy’4, where the demos no longer finds itself in possession of any kratos, then the algorithmicization of this telecracy via the silicon technologies of platform capitalism is already exposing the utter vulnerability of ‘representational’ political systems to a thoroughgoing disintegration at the hands of the ‘owners’ of this data and the manipulators of these algorithms.

This can be described as an ultra-industrial political regression (a new form of what is often called ‘populism’) to which ultra-industrial capitalism tends to give rise. But regardless of the degree to which the leading industrial populists imagine they can cynically keep hold of the reins of power as they exploit the fears and irrationality of the crowd, the enormous risk that they are precipitating is of hubristically engendering processes that will completely run out of all control. All of this is what first began to get going with the shift from an industrial capitalism of production to a hyper-industrial capitalism of consumption a century ago, for which the immensely destructive wars of the twentieth century stand as testament, and it is all this that remains at stake in an ultra-industrial capitalism of algorithmic platforms.

Reinventing economics as the science of counter-entropic struggle in exosomatization

Carbon technologies are thermodynamic: their function is to contribute to the struggle of noetic, technical (or exosomatic) life against its irreducible entropic conditions. But in utilizing these technologies to pursue counter-entropic ends, and given that all negentropic systems are localized systems that are bound to remain entropic in an overall sense, we inevitably produce entropic consequences elsewhere. And when those systems have extended across the entire biosphere, cinesphere, technosphere and exosphere, then this ‘elsewhere’ remains precisely here, and the toxicity they produce is unavoidably self-poisoning, ruining its biospheric element just as does the infusorian in Freud’s petri dish.

Silicon technologies are informational: their function, too, is to contribute to the struggle of exosomatic life against its irreducible entropic conditions. But in this case, the toxicity they produce pollutes not the biosphere but the noetic element of the knowing, technical beings who must nevertheless find the noetic resources to address all of these self-poisonings, whether carbonic or noetic, and to do so by making good collective decisions. It is this division between two kinds of entropic toxicity, and the necessity of recognizing the gravity of informational entropy, that we here seek to highlight.

Most economic theory (like most philosophy) has, to its detriment, remained rooted in a mechanistic physical conception that predates the discovery of the second law of thermodynamics, at least if we believe the economic historian Philip Mirowski5. This means that economic systems are not truly viewed as dynamic processes in perpetual struggle against entropic tendencies but are instead understood as involving one or another kind of static or cyclical equilibrium making possible the fantasy of perpetual growth. From the work of the physicist Erwin Schrödinger, the mathematical biologist Alfred J. Lotka and the economist Nicholas Georgescu-Roegen, however, it becomes possible to see biological (endosomatic) evolution as precisely involving manifold processes amounting to so many struggles against entropy, where these struggles are always localized – at the scale of the cell, the organism, the species, the ecosystem or the biosphere. And it also becomes possible to see that economic processes are what replace these evolutionary tendencies when life becomes technical (exosomatic), still always localized – at the scale of the tribe, ethnic group, society, nation or global economic system.

Mirowski shows that there is a contradiction in neoliberal economics between an absolutized, ‘universal’ conception of ‘the Market’ and a localized (but still informational and computational) conception of specific but highly artificial markets, where the assertion of this universality in fact ends up authorizing the elimination of the wealth of actual knowledge embodied in institutions of exchange of all kinds. His work makes clear that the dangerous turn of recent macroeconomic history – characterized by neoliberalism, financial crisis and proletarianization (in Stiegler’s sense) – has everything to do with the failure of economic theory to incorporate an understanding of entropic and counter-entropic processes, at both the thermodynamic and informational levels. The implicit question it raises is how to reinvent economic theory and practice by incorporating such an understanding from its founding premises.

For a general theory of entropy

This in turn raises the question of the necessity of a theory of general entropy. Such a theory would on the one hand seek to juxtapose and articulate the thermodynamic notion of entropy with the informational notion, and to exceed the limitations especially of the latter. And it would also be in this way an account of the relationship between every kind of counter-entropic system, which is to say every kind of localization and de-localization process that works against the tendency towards the elimination of improbabilities, which is to say the elimination of the past (as what, for any noetic system, opens the possibility of a future). But as Smithson’s association of entropy with both waste and luxury already suggests, this also bears upon Georges Bataille’s ‘notion of expenditure’ and ‘general economy’ (not forgetting that for Bataille, expenditure beyond subsistence is not a question merely of waste but of an irreducible necessity of life).

What this implies, ultimately, is that any such theory is compelled to integrate difficult mathematical, scientific, economic, anthropological and technological questions with others that exceed these divisions between fields of knowledge, in the first place because what is at stake is the counter-entropic function of knowledge itself. Stiegler has indeed begun a project to open up this question of entropy in terms of its thermodynamic, biological, informational and noetic dimensions (in all of its ‘exorganological’ dimensions, in Stiegler’s recent terms), drawing on the work of Vernadsky, Georgescu-Roegen and Lotka, among others, and in discussion with scientists such as Giuseppe Longo, but this requires large-scale transdisciplinary contributory research projects to be established involving scholars across a wide variety of fields. Despite this daunting complexity, such a theory of general entropy has today become a necessity.

In a context in which the globalized systems of consumerist capitalism are reaching their limits, and in the process dragging many other systems past their limits, including geophysical systems such as the climate system, and also including the noetic systems through which alone good collective decisions can ever hope to be made – in such a context, where a cascade of catastrophic system failures seems entirely possible if not highly probable, it is solely on the basis of such a theory that counter-entropic investment prospects with the potential to bifurcate away from such a globally dangerous and monstrous situation can be identified, imagined, invented and realised. Such a bifurcation, and the general theory on which it can be established, will presuppose a reconsideration of the very basis and division of fields of knowledge, but it will also require a complete reorganization of silicon technologies at least as profound as the elimination of carbon technologies called for by the IPCC, and on a comparably short timespan.