We have seen in recent months the blossoming of a protest movement in Weird Facebook, users and memepages wearing pink profile-pictures -- Facebook flags upside-down, signaling S.O.S. -- demanding satisfaction for Zuckerberg’s censorship of their content via automated moderation systems. This “Meme Alliance” engaged in a variety of actions, from complaints to regulatory bodies and Facebook employees to sabotage to a general strike. Though their efforts have served as a coming-out of sorts for meme-scene activists, expanding and strengthening the social graph of users fighting for just self-governance on the Facebook platform, the tactics being used to effect change were, acknowledgedly, ineffective.

First, a series of formal complaints were lodged with the Federal Trade Commission concerning Facebook’s allegedly “fraudulent content review policies and misleading community guidelines.” The circulated call-to-action alleged that Facebook had duplicitously claimed to review reported content while actually relying on an automatic moderation process and that exceptions to the community guidelines were being granted, de facto, to ‘celebrity’ pages which violate the guidelines without censor. The idea seems to have been that Facebook’s guidelines and review processes were meant to be contractual commitments, and that in their violation Facebook was committing fraud.

The problem with this fraud argument is the so-called ‘safe harbor’ provided to hosts of user-generated content by Section 230 of the Communications Decency Act. Section 230 immunizes “interactive computer service providers,” be they social media giants or bloggers with comment sections, from liability for the content users post; at the same time, Section 230 grants these platforms the right to engage in editorial functions, an action which would ordinarily establish that liability (think of how newspapers are responsible for the content which they print). While this provision serves a protective role for smaller entities, providing a defense against, e.g., censorious libel suits for statements made by users, it also grants functionally unlimited control to large social-media platforms over the information which they possess, codifying their right to curative action.

A subsequent change.org petition addressed Mark Zuckerberg and other C-level Facebook executives directly, claiming that automated moderation has caused “psychological distress” for those whose pages had been unpublished and “a loss of revenue” for Facebook as a result of their deletion. While it is certainly true that the faceless destruction of a memepage is a horror -- all that data lost to the void, all that hard work gone in an instant -- the Meme Alliance’s claim that the unpublishing of major memepages results in a loss of revenue for Facebook is dubious. The assumption made was that when a page generates a unit of “reach” (when one of its posts is shown to a user) some quantity of activity is produced and thus a decrease in reach directly affects Facebook’s bottom line. During the later #Zuxit general strike, when pages totalling 20 million likes deactivated their pages for three days, it was noted that their collective reach was in "the hundred millions,” the implication being that this strike would do material damage to Facebook and hopefully provoke concessions. It did not, nor did it even generate a response from the company; regardless, the memers celebrated a well-deserved victory.

The problem with the petition and the #Zuxit strike is that the connection between reach and ads revenue is not as solid as one might think. Assuming that most users utilize the Top Posts sorting option -- the default which Facebook coerces users into using, with “Back to top stories” prominently displayed, but the option to sort-by-newest hidden in a menu -- the loss of any particular post simply results in the inclusion of another in the News Feed. If you take a 400-likes post out of the equation, a 200-likes post still exists to take its place. Users keep scrolling (and viewing ads) in their search for quality content. Though the strike served to provoke an unprecedented level of cooperation among memepages, “work stoppages” like #Zuxit will have little impact on Facebook itself insofar as they do not fundamentally affect the browsing habits of users; the platform’s architecture routes users around such protests.

The strategies employed by the Meme Alliance seem better suited to a different age, when strikes and letters-to-the-editor consumer-movements had a tangible and immediate effect on the targeted organization. As any loss of production or consumption cut into the industrial firm’s revenues, these actions would present themselves as problems -- inefficiencies -- to be solved; in stark contrast, Zuckerberg did not even know that the memepage strike had occurred. This failure of the Meme Alliance’s tactics to register thus demands a reconsideration of the terrain of struggle.


Let us consider a question posed by McKenzie Wark: “What if this is not still capitalism but something worse?” Wark wonders whether today a new economic class has become hegemonic. Perhaps, in passive revolution, the once-dominant capitalist finds himself subordinate to a new economic actor who exploits an emerging set of extractive processes.

Wark claims that the legal concept of property has seen a threefold abstraction. As land transitioned from the hereditary privilege of aristocratic tradition to an exchangeable commodity governed by market forces, a class project emerged via the ownership of the scarce resource of productive space. This “pastoralist” class relied on the abstraction of the legal concept of property to land in their realization of a surplus; with land’s utility expressed as money-rents rather than the “in kind” corvée labor of the feudal system, the value of any productive capacity above that of marginal land could be expropriated from its tenants. Later, as agricultural productivity rose and unneeded farmers streamed into cities, the commodification of manmade infrastructures gave a consistency of method to the capitalist class project; by leveraging its ownership of industrial machinery, capital was able to subject labor-power to market conditions with the effect of alienating the laborer from the product of his efforts and allowing for the extraction of a surplus from the resultant gap. Only recently, however, has the property form found so thorough an abstraction into the “third nature” of information as to provide the terrain for a new mode of production; it is on this extension that the success of the “vectoralist” class rests.

Though Wark’s seminal understanding of vectoralism encompasses a wide range of information commodities -- from patents on GMOs to financial products like exchangeable securities to the just-in-time logistics of WalMart -- one can also map this emerging economic form onto a more restricted set of practices exhibited by Web 2.0 firms. In limiting analysis to this particular subset of the vectoralist form, we can begin to think about the Web in the strategic terms of class conflict. By understanding the particular extractive processes by which these companies profit, the property relations upon which those profits depend, and the laborers from whom they are derived, we can mark the arrangement and orientation of the game pieces in today’s politico-economic conflict -- a fight in and over hyperspace.

With this in mind, Wark’s triplet of “vectors, flows, and stocks” -- three forms of information-as-property which are exploited by the vectoralist class -- are vital to understanding the economic position of social-media platforms.

1. Vectors

Wark uses the concept of a vector to describe means of information communication and manipulation which have “certain definable qualities but … no fixed position.” Two physical examples are infrastructures like roads and telephone lines: they share essential characteristics that determine their use, despite the particularities of each use-case. A digital example might be the TCP/IP protocol which underpins the Internet, routing data from computer to computer by a predefined ruleset but without a fixed path of transit. With a focus on the Web, one can think of how the mechanics of information organization and transmission on social platforms subject the information which traverses them to some predefined system of relations, regardless of the content; i.e. the codeworks which underlie social platforms determine the path and presentation of the information that we post, whatever we post.

This standardization of the movement and transformation of information is, in some sense, the primary way in which value can be extracted from users. Insofar as they control some vector through which information moves, the vectoralist can commodify an element of that medium of information-transit. For example: advertising, the standard way of monetizing a Web-based vector, is the reconfigurement of a website so as to include a scarce resource which can generate a rent. The many lawsuits against the developers of adblocking software illuminate a conflict over where the ownership of a vector ought to end: should a publisher be able to dictate what appears on your screen? Or should their particular vector of information distribution only serve to bring data to a user, to then be manipulated and displayed as the user desires?

Returning to the case of the Meme Alliance, we might consider Facebook’s automated moderation systems to be a vector operating on a temporal axis: although Facebook algorithmically structures whether a post is above or below some other post, the moderation system separately dictates whether a post will persist. It is the Meme Alliance’s tacit claim that Facebook’s vectoral control ought to end at the ordering of posts, not reach so far as to determine their existence -- or, softer, that Facebook’s moderative vector at least be consistent with its written expression in the platform’s community guidelines.

Though these are laudable goals, it is in the interest of the vectoralist to extend vectoral control at every opportunity, not concede it. Examples abound. Apple extended its vectoral control by weaponizing the App Store as a moderative filter, allowing the company to subject the sale of any software for the iPhone series and any in-app microtransactions to a 30% tax. Internet Service Providers have long been throttling torrent traffic identified by deep-packet inspection, repositioning these once-neutral carriers of information as curators in their own right, and they made this same move during their 2014 attack on net neutrality. Even the great GabeN of Valve Software is guilty of this need for extension: the company attempted to subsume game mods, once necessarily held in common due to copyright law, into the monetized vector of the Steam Store.

The fight for vectoral control from within a platform seems a lost cause. These Internet firms seek to always grow, to create or consume vectors such that they become new sources of value. Though there have been successful rejections of attempted extensions -- e.g. Valve quickly cancelled its monetization of mods -- losing any established vector is bad precedent. Undoubtedly, had the Meme Alliance’s petition to the FTC led to anything at all, the whole of Silicon Valley would have stood against it.

2. Flows

Vectoralist platform owners do not only advertise to their users, they also exploit the massive flows of raw data which appear via the infrastructures they control -- users hemorrhage data about themselves and their interrelations. The innovative collection and subsequent use of this data by a vectoral firm can strengthen its product (e.g. the effect of a sorting algorithm) and also its user analytics (e.g. the fidelity of advertising demographics).

Since a better product begets market-share, and since the efficiency of advertising greatly impacts its value, these flows of data have use to a firm even without their active commodification. Companies like Twitter, however, do offer full access to their “firehose” (an apt description of the rate of flow they possess) for a fee, strategically selecting partners who pose no threat to the company’s market position. Other companies closely guard their datasets: Google exploits its vectoral position to produce unique flows of information that are used to reaffirm the company as the leading search engine -- it has no interest in relinquishing that position in exchange for short-term profits.

When the Meme Alliance calls Facebook’s moderation system “automated,” they are referring to its algorithmic reaction to crowdsourced reports by platform users; it “sees” via these reports, with the type and quantity of abuse shaping its conception of individual posts and, by extension, the users and pages who submit them. The flow of these reports provides the empirical data by which Facebook tunes its moderation system, likely taking as its goal a decrease in their number (assumption: fewer reports proves a happier userbase). Though Facebook employs a team to correct false-positives, actually contacting them is near impossible. A Meme Alliance page-admin: “You can never speak to a human being on the phone. You're lucky if you even get an automated response.

Miles Klee of the Daily Dot made the astute point that the frequent moderative interventions in the meme scene are likely “because meme kids report each others' content for a laugh.” The socialization patterns that occur within Weird Facebook are unlike those outside it and whatever understanding the automated moderation system has of how users relate to content is unfit to the new subjectivities of edgy meme-sceners. Insofar as “Report Post” is viewed by these users as simply another UX mechanic -- not unlike the Like or Share or Submit buttons -- Facebook’s algorithmic perception of this Weird subsection of its social graph is radically distorted: it is a bad place full of bad content and bad people.

The Meme Alliance is thus in the position of having to either alter the behavior of the rest of the userbase or at least prove to Facebook that the current model of moderative perception is flawed. In an attempt of the latter, the Meme Alliance brought the fight to Zuckerberg by mass-reporting his posts in an attempt to trigger a post-block or ban -- "#thezuckening" -- but their efforts were unsuccessful. It is unclear whether they simply did not have the numbers needed to overwhelm him, which could be relative to the frequency of a profile’s interactions (Zuckerberg has 80 million friends), or whether they have been right all along that certain profiles and pages are privileged and protected by the platform. Regardless, the flows of data which shape the platform’s moderative vector will likely drown the meme-scene long before being overcome.

3. Stocks

The vectoralist class-project is unified by the desire to strengthen the legislative defense of the information property-form; this drive gives vectoralists coherence as a historical bloc. The importance of “intellectual property” stems in part from the fact that vectors of information and datasets siphoned from flows of information require legal protection: even Google is just one complete data-leak away from being usurped as the King of Search. Indeed, nearly every social platform can be considered “software as a service,” with only its blackboxed algorithms and proprietary indexes preventing it from being owned in common.

In contrast, some firms use this relation to information-as-property to generate value, e.g. “data lockers” like Netflix and Spotify which use different methods to monetize the vast stocks of information which they hold. While Spotify’s advertisements might be considered the commodification of a vector rather than a stock of information, the use of their platform is immediately predicated upon the illegality of alternative vectors of access to the music which they host. The commodification of stocks of information is thus about keeping nonrivalrous resources scarce via the threat of state violence, despite the ontology of information implying the possibility of its ubiquitous use and possession.

It was inevitable that the FTC complaints would prove impotent. The control of stocks of information is perhaps the essential shared interest of the vectoralist class, and the ability to determine, at will, what information is made available when, where, and to whom is the very capacity by which the vectoralist profits; to challenge the curative capacities protected by CDA Section 230 is to challenge the very sovereignty of platforms. It was also inevitable that the petition to Zuckerberg and the subsequent #Zuxit strike would fail to reach Zuckerberg’s desk, ensured by the qualitative differences between the capitalist model which relies on the production and consumption of commodities and the vectoralist model which profits by the movement of users through a monetized hyperspace. So long as there is the same amount of engagement with the vector, a platform owner can care less about the stocks of information which generate it; it is irrelevant how much content is produced, let alone of what type. The strike’s success required that memepages were exclusively responsible for a large portion of Facebook’s traffic, an ask too great for any page on the platform, Weird or not.

Sidebar: Class Competition

It is important to understand that though these three methods of value extraction beget a particular set of shared interests and thus direct the actions of vectoralists as a political bloc, they should in no way be understood as a kind of monolithic force. Just as the capitalists before them, this newly hegemonic class engages in a certain infighting -- e.g. the patent wars of recent years -- as they jockey for power on this burgeoning economic terrain. It is equally important to understand that the emergence of a vectoralist class does not suggest the disappearance of previous economic forms. Not only do capitalists and pastoralists still exist, they actively fight for advantage.

For our purposes, however, the Web’s existence as “third nature” excludes these classes from economic leadership. Though manufacturers can indeed sell products from Web stores on their homepages, their dependence upon social-media vectors for advertising, as well as the ubiquitous listing of their products on larger digital markets like Amazon and AliBaba, suggests a certain subordination of industrial capital. Labor, too, finds itself in a precarious position, with digital labor-markets, e.g. Amazon’s Mechanical Turk, Uber or TaskRabbit, dictating the terms of Web-mediated employment. Pastoralists are obliterated by the potentially infinite “land” of hyperspace -- though ISPs and cloud-storage providers seek to insert themselves as vectoral-age landlord equivalents. It is important to say it: any class can be dropped from the equation until that time at which it begins to fight vectorally.


Wark describes a universal antagonist class which resists the extractive processes of vectoralism: it is the “hacker” class which, in its creation and commonization of new information and new ways of making information, fights for the user. Though not a replacement for Marx’s proletarian, Wark does position the hacker as having a unique ability to challenge the vectoralist on this new terrain.

In one sense we are all hackers, abstracting the real from the virtual in the production of new information (which always also contains the possibility of becoming a new practice of information production). Yet, in the Web 2.0 era, our digital lives are being scraped, with even rote activity data-mined to its limits and then sold off to large players in other markets. There is no action we perform on a private vector which, as productive labor, cannot be transmuted into exchange value.

Our productions are reduced to new “universal equivalents" -- views, likes, retweets, et al. -- and we are even encouraged to join the vectoralist game: in an effort to increase the user-metrics of vectors they control, some vectoralist firms are willing to give back to us a portion of the value generated by the content we produce. It is in this way that the vectoralist class (falsely) proposes itself as universal, with every production sourced back to its creator and every curator or signal-booster credited (a percentage) for their efforts. The hacker class becomes coherent only when it is generally understood that the vectoralist model forces each and every one of us to buy back the information which would otherwise be held in common. This is even and especially true when we would receive a cut of revenues in exchange for our labor.

In another sense, however, those who produce the infrastructures by which we communicate -- both physical and memetic -- are the “true” hackers who find their work torn from them by the vectoralist class. “Steve Jobs didn’t invent the iPhone,” nor did Zuckerberg create Facebook as we know it today. (The Meme Alliance’s elevation of “Zucc” to a panoptic Big Brother is, of course, in jest.) Though the free software movement has provided the ideological groundwork for communal creation, too often high salaries, free drinks, and laundry services win over potential hacker-slackers who would have otherwise lived cheap and coded hard. As a result, an incalculable effort has been spent solidifying the market position of social-media “walled gardens.”

Fortunately, today, the means of vector production have seen a radical democratization. A knowledge gap exists, as does the digital divide, yet there is the distinct possibility of a small group of individuals creating a competitive vector which could propose itself, at the level of code, as an “information commons.”

An Analysis of the Situation

It is apparent that the development of new forces of production have provided a basis for the emergence of two new social classes, but the asymmetry in their class cohesion is striking. The hacker class has never come close to establishing a consciousness comparable to that of the vectoralists -- not during the encryption wars, not at the height of the Pirate Bay file-sharing era, not today. There are cyber-radicals, of course, and the free software movement is important, yet, in terms of society at large, there is scant organization in favor of the freedom of information and the rejection of its property form.

There are, however, a number of artisan classes cohering under vectoralist hegemony. “YouTubers” refer to themselves as a such when railing against Google’s content and DMCA policies. “Gamers” have oriented themselves against the vectoral organization of pay-to-play games journalism (and been captured, in large part, by the alt-right). Reddit moderators revolted via “blackout” last year. This suggests that the hacker class has reached a certain stage wherein class identification occurs according to membership in a group which is conscious of its unity and homogeneity, and of the need to organize it, but with a blindness to the case of the wider social group of Internet users. The radicalization and unification of these groups, and the active prevention of their assimilation into some kind of “petit vectoralist” creative class -- “vectoral” in the sense that they profit by flows of views, likes, and subscribes which their stocks of content generate; “petit” insofar as they do not own the vectors upon which they labor -- is paramount.

Tactical Reorientation

While the Meme Alliance represents a coherent attempt to build consciousness within the bounds of their platform and productive grouping, there is a need to expand the object of their critique. By understanding their demands as singular, as particular to the Facebook platform, the company’s motivations remain distorted.

The complaints to the FTC were made with the understanding that Facebook was fraudulently unpublishing memepages, that its moderative activities were not consistent with the platform’s published community guidelines; insofar as content creators rely on Facebook for their livelihoods, arbitrary deletions harm the very content creators who stock Facebook with content to monetize. How could someone knowingly take on that risk to produce full-time on and for the platform? It is a fair question -- fair enough that it is currently being asked all over the Web. Youtubers, for example, are in near constant revolt over YouTube’s copyright complaint system which is abused regularly by corporations and individuals alike, often leading to producers’ ruin. Twitch inconsistently polices the morality of both its users and the games they stream. Twitter seems to ban users based on the media attention that they attract. Reddit has been quarantining the worst of its subreddits without a consistency of application.

The idea, then, that the protest of Facebook’s curation of content on its platform would bring about redress is fundamentally flawed. The assumption was that the inconsistency of policy and practice was accidental or somehow overlooked but, in fact, it is a “feature” implemented by every market leader: on nearly every monetized platform, users labor as subjects of an opaque, Kafkaesque system of governance.

The motivation is apparent. Deviant content and its concomitant communities tarnish the reputation of these platforms, decreasing competition for the space carved out for advertisements. In 2013, when 4chan’s self-serve advertising program launched, the cost per 1,000 impressions (CPM) on the site was $0.20; in the same year, Facebook’s desktop CPM averaged $0.75 and its mobile CPM an incredible $5.31. There is a reason 4chan is going broke and not just because it serves an ad-blocking community: advertisements on edgy platforms are just not worth much. The Meme Alliance is asking for nothing less than sovereignty -- “Facebook must allow content creators the option to turn off the unreliable automated moderation, which frequently fails to distinguish between terms-violating content and harmless content” -- but the platform cannot and will not grant it.

If we want to escape the iron heel of Zucc and his ilk then we must fight on the vectoral terrain of the 21st century. Without the production of an alternative, without an alternative held in common, one vectoralist or another will always seek the maximum surplus from our socialization -- OC be damned. That means the smoothing over of rough edges to create safe space for advertisers and a governance policy which understands content not in terms of dankness, but by the quantity of reports it receives.

The Philosopher’s Meme prototyped an oekaki imageboard which organized content by flavor, ensuring spicy memes could accurately be distinguished from salty and sour. A friend made an image archive, co-curated by user-submitted tags, which could be fed into a Facebook chatbot and thus escape censor. Kiwi anarchists are imagining new ways to order information which might provide a hyperspatial structure for new kinds of memery altogether. If the cumulative reach of the meme-scene is as great as we say, are we not a bullhorn? Could we not seed with users any number of projects which hope to provide a new terrain for memetic play?

Keep your pages and profiles -- information distribution networks are a tactical necessity -- but expect no more from them. Prevent censorship by circulating the URLs of backup pages and alerts of unpublishings, but prepare to get zucked. Always post fresh content, growing your reach, but make somewhere else home. This is not a call for market-competition or exodus. The value of Facebook’s 1.6-billion-user network is too great to overcome in direct competition -- do you know anyone who used ello for more than a week? -- and only a fool believes the meme-scene would delete their accounts en masse, relinquishing the very social utility that makes Weird Facebook a vibrant scene in the first place. This is, however, a demand that we reclaim an outside. They say, “Facebook is eating the Web,” but we do not have to let it.

Facebook can only ever be an advertising platform; make it work for you instead of Zuck. Insofar as the avantguard meme scene has often provided the primitive accumulation of users for social-media platforms, our principal project can and must be the construction of alternatives.