A recent paper by Hamid Ekbia presents an interesting Marxian theory of the relation between exploitation and computer networks. The paper is intended as an intervention in to discussions of the accumulation of value in what is now called cognitive capitalism (I’ve attempted to synthesize some of that literature here). The most interesting part of the Ekbia’s paper seems to me that he’s able to construct a coherent notion of class (or close to class – he acknowledges that it’s not quite a class in the strict Marxian sense) within those who are part of the networked economy. In particular, he is able to locate those who are exploited and to roughly define them as a group: the “condensers.” The problem of locating a specific exploited class is important and salient partly because there is no way for a Marxian theory of value to work unless somebody is exploited, but also because the behavior of prosumers in particular has been the subject of intense controversy, particularly on the subject of whether they produce value. Ekbia’s contribution, it seems to me, is to show how and why some prosumers manage to be exploited.
Ekbia’s thesis depends on rewriting the notion of class along the lines of network inclusion: “Digital inclusion – in the sense of being connected to a network, not being a member of the privileged class – has become the modus operandi of current capitalism” (168). That is, current information capital operates through networks. This causes difficulty in determining where value comes from, as it is not, for example, obviously connected to socially average labor time. One example is the creation of network value. Compared to, say, a single car, a single telephone is pretty worthless, but the more phones that are connected to the network, the more value the network as a whole realizes, both to those who use the network and to those who own it. As the market valuation of companies like Facebook suggests, the bulk of that value goes to those who own the networks. A second tier of user is able to exploit the uneven distribution of connections across the network, extracting value by way of being well-connected to other points on the network. Those who can make a living monetizing their YouTube videos come to mind here. In both cases, the revenue stream importantly depends on the ability to sell advertising, on the premise that such targeted advertising will reach consumers whose preferences have been accurately identified: advertising will become efficient. It is this dynamic that leads me to think that the entire system depends on an originary accumulation of users’ preferences on the part of the system (this may be purely aspirational, and the model may fail), a point to which I will return in a moment.
For Ekbia, there are two classes of those who profit from the labor. In both cases, the system values individuals by their network location; “this is a world where forging relationships can be a source of profit, and where people’s value derives from their degree of ‘mobility’ – that feature of the connectionist world most relevant to the mechanisms of value extraction that are particular to this world” (169). The two groups are “network extenders” – those who forge new connections, etc. – and “opportunistic networkers” – those who utilize their own position in the network to strategic advantage (the YouTube users who monetize their videos, for example). Both depend on the condensors, i.e., those who are “minimally mobile or involved in a kind of mobility that is ineffective in leveraging their personal advantage” (169). More specifically, “the labor and efforts of condensers are appropriated by extenders and networkers, not on the basis of class relationship, but through their limited access to network’s social, material, and informational resources” (170) and:
“Condensation, as the generic mechanism of value extraction in the connexionist world, consists of those activities that are needed to keep the otherwise sparse connections of networks live, populated, operational, and productive. Condensers are not members of the same class; the degree of heterogeneity in their relationship to means of production makes it unwarranted to include them all in a single Marxian class” (170).
So not quite a Marxian class, but the permanently marginal (he later compares them to Marx’s surplus labor army), whose relative lack of power enables their exploitation. It seems to me that the difficulty in calling condensors a class in the strict sense actually points to their condition on the margins of the system of capital, which means that their exploitation needs to be (also) thought as a matter of primitive accumulation. Tracking Marx’s work on primitive accumulation (and Harvey’s use of it in terms of “accumulation by dispossession”), one sees precisely the heterogeneity of those who are eventually brought into the system, and the heterogeneous ways they are incorporated, whether through the loss of cultural values to intellectual property, through biopiracy, or through Marx’s original example of the loss of common spaces for grazing through enclosure laws.
But what does this mean? I would like to suggest that a part of their value is in their putative ability to stabilize markets and provide a measure for network value, and that this function points us in the direction of primitive accumulation. Among the autonomists, Negri in particular has repeatedly suggested that current capital has lost its measure of value (interestingly, this narrative goes at least as far back as his Job book, a story he interprets as being about a loss of value and proportionality). As Negri puts it in a 1999 article, “measuring labor, and thus ordering it and leading it back to a theory of value, is impossible when, as today, labor-power is no longer either outside or inside capitalist command” (80). Ekbia reaches a similar point with reference to Foucault’s Discipline and Punish; the factory worker is no longer the ideal subject. Instead, “the archetype of the ideal subject now is the teenage gamer, the obsessed social media member, and the permanently reskilled, reeducated, and mobile professional” (172). Their participation in the system returns some use value to them, but the greater value is in the system’s ability to mine their preferences in order to get them to part with money. In this movement, they become subjects of capital.
This is done through a variety of mechanisms; I will just indicate two candidate examples. First, an article in Slate today suggests that Twitter is developing complex algorithms to optimize a user’s feed based on the system’s understanding of what kinds of tweets its users want to see (for example, they might want to see tweets from users with more followers than less, and they might want to see frequently retweeted comments by higher status users over those less retweeted). The site claims that this tweaking has already made it more sticky (users stay on longer) and increased the posting activity of users. This sort of data analytics is reading user preferences according to an algorithmic process that is opaque to them, and constructing the site’s ability to sell advertising. For a second example, consider work on Facebook likes. It is possible to infer a vast amount of information from a user’s Facebook likes, including personality type, which is predictive of future behavior. The translation from likes to preferences and aspects of identity is important because it is entirely based on metadata: the user does not actually have to say anything about a given product or personal characteristic for the site to infer what she would likely think about it.
Again, the drive behind all this analysis seems to be to get the user to signal what her commodity preferences are, in order to make advertising more effective (it is also about branding and forming/rewarding brand attachments, but developing that point will add too much complication here). The user’s privacy, in the sense of preferences shielded from the market, are systematically handed over to the sites whether the user wants this to happen or not, and this extraction of privacy makes the user more vulnerable to advertising and more vulnerable to persuasion and formation as a consumer of relevant products. That is, the information asymmetry between users and sites is increased, and users are increasingly subject to manipulation. Some of this manipulation is toward consuming specific products; other manipulation is in the foreclosure of non-market ways of relating to others through the sites. As I suggested elsewhere, there is a striking coincidence in Marx’s own writing between primitive value extraction and what we would now identify (via Foucault) as processes of subjectification – the language used is very similar, as are the processes described.
This account of course does not account for the value generated by the exploitation of those who are outside these networks and in more traditional capitalist relations, whether as factory workers, agricultural workers, and so on. Nor does it account for the oftentimes violent coercion exerted to get those individuals into the apparatus of cognitive capital. However, once they are in there (voluntarily or not), it offers a plausible way to explain how the majority are exploited while a few are able to reap the benefits of that exploitation.