See original here.
For some time now, a façade of techno-optimism has obscured the political reality that technology is not neutral. While Marx saw both oppressive and liberatory potential in technological systems, recent narratives have tended to neglect how technology has historically been used to deepen exploitation, rather than to overcome it. In his new book, Breaking Things At Work, Gavin Mueller reclaims Marx as a “cartographer of proletarian struggle” who recognises that “[t]he struggles against machines were the struggles against the society that utilised them…”. 
Mueller is clear throughout the work on the way in which value becomes bound up with and is weaponised through machinery. Marx observed that machinery revolutionises, “the agency through which the capital-labour relation is formally mediated, i.e. the contract between the worker and the capitalist”.  Mueller makes a similar point a few times in the book, first when he introduces the notion of subsumption or the ways in which human activity, particularly labour, is reorganised according to imperatives that serve the logic of capital, rather than society. Later in the book he applies this framework to the introduction of the internet and the digitalisation of everyday life, claiming that the early internet was not yet fully subordinated to capital and it was only with the more recent datafied form that the internet and digital technologies became fully subordinate to capital. In order, however, to understand the implications of the more recent technological changes for working life, it is crucial we understand how value governs these relations and how the imperative to extract surplus-value weaponises the machine against workers.
The notion of subsumption is key to understanding capitalism because it entails the standardisation of the labour process. Along with the creation of liberalised markets and competition, it is one of the preconditions for the emergence of capitalist relations. There are two levels of subsumption. The first level Marx refers to as ‘formal’ subsumption, which forces producers to deliver surplus labour for capitalists, rather than for themselves. This occurs as capital comes to own and control the means of production, rather than labourers. Such a process took place in the working practices of craftsmen, such as weavers, during the time of the Luddites, but remains an active part of the labour process today, when informal or self-employed domestic workers are compelled to work for a platform like Care.com. The second level of subsumption is more substantial, as it completely revolutionises a production process and the productivity of the workers through introducing new technology and reorganising the labour process. The classic example of this is the movement from simple cooperation to manufacturing and large-scale industry, which required a detail division of labour.
In Capital, Marx drew on insights from Charles Babbage’ On the Economy of Machinery and Manufactures (1832) and Andrew Ure’s The Philosophy of Manufactures (1835) to show how the disaggregation of the production process into precise components allowed for both the most efficient use of labour and the exercise of power over workers. Marx understood that the process of subsumption was intensely political. In Labour and Monopoly Capital (1974) Harry Braverman later argued that managers are structurally incentivised to despotically substitute simple for complex labour and constant capital (machinery) for variable capital (labour-power). Breaking down complex and highly skilled labour processes into ever simpler, measurable components is a precondition for the mechanisation and automation of that labour. It allows for more value to be squeezed out of every minute of the day. To understand how this works, we need to understand how value and machinery relate to one another.
Most people familiar with Marx are familiar with the labour theory of value. Very simply, the labour theory of value holds that surplus-value is the source of profit under capitalism. In order to produce said surplus, firms require workers to work longer than the time required to simply reproduce their lives (the value of their labour-power). This surplus-labour-time is unpaid and forms the substance of surplus-value. The introduction of new machinery tends to reduce the cost of the product, since such machinery needs less labour to produce the same amount of product. Reductions in the prices of market products means that the costs of the social reproduction of the workforce also lessen, meaning the cost of labour. New machinery increases both fixed capital (through machinery itself), but also the overall constant capital by allowing a greater amount of raw materials to be transformed by labour-power into commodities. This can exponentially increase the productivity of labour and the efficiency of the use of raw materials, while keeping wages low.
According to the LTV, machinery creates no new value, but instead “yields up its own value to the product it serves to beget”.  The product is “made dearer in proportion to the value of the machine”. The amount of value transferred by the machine to the product depends on the total value of machinery and the rate of transfer – the less labour it took to make the machinery, the less value it can transfer. Paradoxically, the “less value it gives up” over the long-term, the more productive it is.  The value contribution of machinery is multiplied in a complex system where the entire machine framework is “consumed in common by its numerous working parts” including the energy source, transmitting mechanisms, etc. by all the other operating machines in the network.  Thus the more productive a machine is, the “more its services approach those rendered by natural forces”. 
However, the productivity-enhancing effects of machines do not continue indefinitely. There are countervailing forces such as competition, which drive down profit rates within industries and can have knock-on effects in adjacent industries and throughout the value-chain. Other countervailing tendencies stem specifically from technological change, especially in the age of datafication and AI. This is because the value a machine contributes is not simply due to physical deterioration that is more or less proportional to its use and value-transfer. There is also a second kind of depreciation that concerns a loss in exchange-value, which Marx refers to as “moral depreciation”. Such depreciation of a machine occurs because machines of the same kind are being produced more cheaply or because better machines are entering into competition with it, or what is more often the case, a combination of both. In this situation of technological obsolescence, the value of the machine is “no longer determined by the necessary labour-time actually objectified in it, but by the labour-time necessary to reproduce either it or the better machine”. 
New technologies enabled by AI and datafication, which we can also understand as virtual machines, are introducing new ways of substituting machine power for human labour power. Traditionally, automation tends to facilitate deskilling in existing labour processes while increasing the monitoring and surveillance of workers, resulting in less worker discretion and lowering the quality of work. Indeed, many studies of new forms of automation show this already to be the case. 
At the same time, AI and datafication are transforming the value chain for both goods and services, as intangible assets gain importance. Intangible assets include things such as branding, business processes, databases, and algorithms and other forms of intellectual property. Intangible assets typically require a significant capital investment initially for research and development, but then take less labour to reproduce through maintenance and updates. For software, prices of production are typically 1.5 times the average cost of wages for software occupations.  The low cost of reproduction means that these products, particularly those that rely primarily on intellectual property, to continue to be profitable, must become sources of rent.  Companies now extract or expropriate data from the entire population and transform it into an intangible asset or raw material for the reproduction of virtual machinery via machine-learning AI. This is why data is now understood as a form of capital. The connected vehicles that UBER drivers use, linked to apps on their smartphones, each produce valuable datasets. Some of the largest companies in the world like Microsoft and Google are now essentially a network of virtual machines built on data capital.
Platforms accumulate monopoly rents through their X-as-a-service model, yet also accumulate data assets of a more speculative quality that can be leveraged for future valuations through financial meta-platforms. For example, many white goods like washing machines used to be relatively undifferentiated products. Price and brand recognition were the primary basis of competition and market stratification. Laundry detergent was a key differentiating input and the intellectual property of its chemical composition was a valuable (though static) intangible asset. In new “smart” washing machines, connected through the internet of things, sensors collect a wealth of data, which can be then leveraged for product enhancements, consumer research, and the expansion into other value-added services.  The collecting of this data becomes a continual process of expansion and revision of intangible assets as value is captured both within the labour process and outside of it.
The datafication of social relations (both in work and beyond it) is assetisation in motion. Similar to the way that the craft skills of weavers were made redundant by the introduction of power looms, the activities of millions of people both within and beyond the workplace are being transformed into intangible assets as part of virtual machinery. With the assetisation of data, a dual value extraction is now possible that not only enables a digital Taylorism that contributes to excessive surveillance, stress and other negative consequences for workers, but also an additional dimension of exploitation.
In Breaking Things at Work, Mueller reminds us that the new antagonism between consumer and platform over data capture is not unlike the struggle between worker and capitalist over wages and the working day. In submitting to the collection of our data as consumers and workers, we engage in an exchange without a bargain. Virtual machines are catalysing the transition to a digitally-dystopian stage of rentier capitalism. Yet, from hacking to piracy and strikes, there are many modes of resistance to be pursued. To avoid the digital deluge, we must look beyond techno-fetishism, recognise the exploitation inherent in capitalism’s approach to technology, and break the virtual machines.
 Mueller, G., 2021. Breaking Things at Work: The Luddites Are Right About Why You Hate Your Job. Verso, London. p. 24
 Marx, K., 1976. Capital, Volume I: A critique of Political Economy. Penguin in association with New Left Review, Harmondsworth. p. 519
 Marx, K., 1976. Capital, Volume I: A critique of Political Economy. Penguin in association with New Left Review, Harmondsworth. p. 509
 Ibid. p. 512
 Ibid. p. 510
 Ibid. p. 512
 Ibid. p. 528
 See Spencer, D., Cole, M., Joyce, S., Whittaker, X., Stuart, M., n.d. Digital automation and the future of work, Panel for the Future of Science and Technology. European Parliament, Brussels. https://www.europarl.europa.eu/stoa/en/document/EPRS_STU(2021)656311
 See Haskel, J., Westlake, S., 2017. Capitalism without capital the rise of the intangible economy. Princeton University Press. p. 47 https://doi.org/10.1515/9781400888320
Heading 1 Teixeira, R.A., Rotta, T.N., 2012. Valueless Knowledge-Commodities and Financialization: Productive and Financial Dimensions of Capital Autonomization. Review of Radical Political Economics 44, 448–467. https://doi.org/10.1177/0486613411434387
 Weber, Steven. ‘Data, Development, and Growth’. Business and Politics, vol. 19, no. 3, 2017, pp. 1–27, doi:10.1017/bap.2017.3, p. 419.
Dr Matt Cole is a Postdoctoral Researcher at the Oxford Internet Institute, University of Oxford where he studies AI, datafication and work in global platform economy. He recently co-authored a report on ‘Digital Automation and the Future of Work’ for the European Parliament and has written for the Socialist Register, Tribune, Salvage and Novara Media among others.
This is Part 2 of the Verso Roundtable on Breaking Things At Work. Follow the link for the other articles in the series.