My Life in the Cloud

What’s the internet, and where do we belong in it? A postscript on “I’m that angel,” the author’s book and performance for the world’s data centers.

Sep 22 2015, 3:20pm

Photo: Inside Suite/Wikimedia Commons

A year after graduating from college, I was living in London and having my first taste of precarity. Almost every day, I'd troll the Craigslist job postings, quickly learning that my standards were much lower than I thought. I moved other people's objects around the city, stood in rooms holding canapés, stood in alleys holding trash bags filled with canapés. The only time I drew the line was when my catering company asked me to stand in rooms holding canapés in a diaper and bonnet. True story.

I also moved words around Microsoft: for forgotten art journals, Jewish cultural centers, bottom-feeding media blogs. A house cleaning company hired me to write original content for its website. Over a few months, I churned out posts about kitchen cleaning tips, dust mite prevention, and the best ways for young mothers to organize their daily tasks (writing from great personal experience, of course).

Some farmers just burn out, pure and simple.

It would have been a stretch to call myself a creative, let alone a brain-for-hire. Across all of these jobs, whether on the web or in the city, I rarely felt like I was doing more than exercising my hands: moving digits to move things—aesthetic things, informational things. Even before I had the critical vocabulary to articulate this feeling, I could recognize the reach of it. I was merely one of many actors playing out a new labor paradigm.


That was 2006, roughly the time when the first great content farms came into being—when my scattershot online content production formalized itself into an industry proper.

So, what is a content farm? Simply put, a content farm is an online news organization that generates articles based on trending topics.

Take Associated Content, created in 2005 and later rebranded as Yahoo Voices! In its heyday, this farm used robots to scan websites like Google Trends and generate relevant article prompts, which could be claimed and written by its content farmers, sometimes with as short a turnaround as thirty minutes. Associated Content was known, at its peak, to publish 10,000 pieces of content a week.

While it's true that there's a human on either end of content-farmed articles, we should not mistake their economic purpose: they're being written to game search-engine algorithms. They deploy trending language to ensure higher rankings. The articles get more clicks, their advertisers more "eyeballs," to use industry parlance. Every eyeball, another few cents, and so on and so forth.

Screenshot: Demand Media

As another example, consider Demand Media, a consortium of websites like eHow and Livestrong. In 2009, the company was projected to be publishing one million items a month. Rumor had it Demand was so profitable that major news hubs like The New York Times would soon go out of business. As of 2010, Demand was the 17th largest web property in the US—and the first content farm to go public.

Now, content farms have been criticized for lessening the quality of platforms like Google, bringing users to lesser-quality sites. They've been described as the fast food chains of the information superhighway, and the reputations of search engines have suffered by proxy. One response came in the form of Google Panda, a more robust algorithm that the company introduced in 2011. Panda was designed to lower the search rank of thin, low-quality sites, thus directly targeting the stuff coming out of the farms. Since then, content farms have continued to spread, but in general, such measures have had substantial effects on the industry.

Screengrab: YouTube

Demand Media's sites, for instance, lost roughly a quarter of their traffic in the immediate wake of Panda; moreover, the company's stock crashed and stayed down. Yahoo Voices! shut down in the summer of 2014, sending regular contributors scrambling. The era of the big content farms is coming to a close, and by many accounts, we are transitioning to a "clickbait" economy.


To come full circle: I began as a precarious content generator at the same time as the rise of these farms, and around 2011, when I returned to research them, I found quite a bit of scorched earth.

Nonetheless, I wanted to think about the content farmer, this member of the creative class in pursuit of the self-liberation and entrepreneurial fame promised by post-Fordist rhetoric—who believes in Millennial myths, despite all evidence to the contrary.

I've spent time in online forums with farmers, where I've heard some common stories. Most also work part- or full-time jobs, but are keen to pursue their true passion of writing. Content farms thus provide room to hone their skills and generate clips; few aspiring writers view them as a primary revenue stream. Demand Media happens to be one of the best remunerators in the field, paying around $15 an article. Several other farms pay an entry-level fee of a cent per word, which can increase to a few cents after a certain number of contributions.

Content farms, in short, are a means to a different end, yet many of the farmers write for companies that don't provide bylines, preventing writers from using their articles as clips. Additionally, given the trying times the industry is facing, farms have been known to close sites and slash pay without warning. Their farmers can either consent to the changes or leave.

Some farmers just burn out, pure and simple. Others attempt to take action.

To give an example: In 2011, AOL bought the Huffington Post for $315 million, turning the political blog into an advertising-financed operation. Jonathan Tasini, who had contributed 216 unpaid articles to the blog, filed a $105 million class action suit on behalf of himself and 9,000 other unpaid bloggers, arguing they deserved a cut.

The case ended up being dismissed on the grounds that these writers had never worked with the expectation of being paid; their argument, in short, was "baseless." Huffington gave a similar defense, adding that personal passion and cultural capital were sufficient motivators for her contributors.

Many of us experience similar labor conditions in our online and offline lives.

There are other cases worth discussion, aimed at companies like Yelp and Crowdflower. Each is negotiating the terms of a specific web platform, yet they all raise a basic question about the labor rights of digital content generators—a question that certainly extends to the farmers.

Of course, many of us experience similar labor conditions in our online and offline lives. The content farmer, as such, is not an exceptional case, but an exemplary one for considering different writing on the informational age: from Jodi Dean's theory of "communicative capitalism"—where messages are made to be circulated, not generate dialogue—to Michael Hardt's claim that the affective and emotional dimensions of human relations have assumed dominant positions within the contemporary economy.

As an instrument, affect has been used to naturalize emerging forms of labor and sociality, giving the figures but hiding the ground. The content farm is one site where we see this in action: Articles are pitched to "populist" interest, empowering their writers and edifying readers with...listicles and how-to guides.

The thin, necessary veneer we maintain is that we are humans—are welcomed into cyberspace in our full capacity as humans. True, this may all be a front for a game of economics played by robots and search-engines and advertisers, but if we have a role, let it be more than an "eyeball." Let our online participation deliver us into a future that's more than monetizable. We know we're "eyeballs," but still.


In one of his cattiest moments, American novelist Truman Capote said of the work of Jack Kerouac, "That's not writing, that's typing." What would Capote have made of writing on the internet? The content farmer also interests me for this reason: for what we do upon recognizing that we're writing less for humans than algorithmic capture—that our words will mean more to the Google robot than to anyone else.

To American conceptual poet Kenneth Goldsmith, we are quickly adapting to the currents of the online world, growing less and less like the readers of language than the parsers of text. In attempting to keep up with the informational demand, we sort through content much like a robot reader.

In the not-too-distant future, according to Goldsmith's account, we give up this John Henry narrative. We are already writing more for machines than humans—and our machines are also doing their own farming and writing. So why not just let them do all the work?

This is a question Canadian poet Christian Bök directs to the field of poetry, asking, "Why hire a poet to write a poem when the poem can in fact write itself?" These poems of the future, to Goldsmith, "might be best not to read at all; they may be better to think about."

Here we have a familiar endgame. Whether we are dealing with poetic, artistic or other labor, the theory goes, technology can decrease our working time and increase the time for personal reflection. It makes a gamble on the future: that we will benefit from ceding more control to machines. Yet somewhere in these quotes is a darker intimation. Perhaps we won't have much choice in the matter.


With these thoughts in mind, I created I'm that angel, a book made to be performed in data centers. My project is very much a response to this endgame, departing from the belief that we should not just make do with our online roles, nor abandon them wholesale, but put them in the service of imagining new forms of subjectivity and sociality.

The book is narrated from the perspective of a content farmer in the throes of intense semiotic anxiety. At the time I began writing it, "the cloud" was entering common parlance, and I frequently had the experience of hearing friends use the term, only to then remark, "I don't know what it means, but I know that we're supposed to use it."

Data center at Virginia Tech. Photo: Christopher Brown/Flickr

I've lived through many generations of digital technology, but this transition stood out, perhaps for how a diffuse, natural form could become a diffuse, naturalized metaphor—for how a symbol could come to operate without needing to clarify what, in fact, it stood for.

Herein lies the semiotic anxiety of my protagonist: not understanding what the cloud signifies, and thus how he can come to relate to it. If he knows himself as a body of certain extensivity, and if the cloud is diffuse, how does he know if he's already inside or outside of it? And if, to paraphrase Groucho Marx, the cloud is a club that's soliciting his membership, would he even want to be a part of it?

These are ridiculous questions, of course—little psychic wisps that cloud the content farmer living within the cloud. We commonly describe the stultifying effects of such anxiety: doubt, worry, the onset of paralysis. Yet in my book, anxiety is the motor that winds the protagonist up, spins him around, and thereby confirms that for all of his online perambulations, he also holds a place and a body in the world.

Content farming establishes two parameters: a necessary quota of trending language, an article of sufficient length and quality. So I wondered: with limited editorial oversight, to what other use could a content farmed article be put? Could one claim the linguistic space around quotas, the spans between buzzwords? Could a memoir be written by a farmer smuggling the self into industrialized writing?

In his 1917 essay, "Art as Technique," the Russian formalist Viktor Shklovskiĭ called for strategies to work against the "automatism of perception" and to "defamiliarize" the habituated aspects of our world. "[A]rt exists that one may recover the sensation of life," he writes. "[I]t exists to make one feel things, to make the stone stony."

I'm that angel is an attempt to make a stony stone—a scatological stone. The body is in protest against our diffuse, master metaphor, and language turns thick and suspicious.

Here's an excerpt:


This would be a good point to talk about the physical internet. But first: Addressing the 2005 graduating class of Kenyon College, David Foster Wallace tells the following joke:

"There are these two young fish swimming along and they happen to meet an older fish swimming the other way, who nods at them and says 'Morning, boys. How's the water?' And the two young fish swim on for a bit, and then eventually one of them looks over at the other and goes, "What the hell is water?"

So my protagonist, as I mentioned, suffers from semiotic anxiety. He doesn't understand what the cloud symbolizes and thus can't meaningfully relate to it. Some of us may have once suffered similar anxiety, though we now know what the cloud signifies: the thin devices and remote data storage underpinned by the massive infrastructure of the physical internet. It's an inaccurate metaphor, as our cloud is tethered to the earth like a balloon, though it's understandable why the internet was made into a heavenly body and not a flimsy piece of manmade rubber.

When I began writing I'm that angel in 2011, I felt compelled to demystify the cloud, to pin it down and open it up, to stage a re-encounter between oneself and one's data—to chant, in reminder: "This is water, this is water."

To date, I've held events in more than ten data centers in Europe and North America, inviting small audiences for readings and facilities tours, conducted by data center personnel. These events change with every venue, with the given company, and with what guests bring to the conversation. Allow me to share a few notable stops:

1. The Pionen White Mountains facility in Stockholm, located in a former civil defense center. The name is somewhat of an exaggeration, as Pionen is not in the mountains but beneath an elevated park on a residential street—far from secretive, far from remote.

Anyways, this site is famous for two things. One, when it was turned into a data center in 2008, the company decided to trick it out James Bond style. Two, if you're in need of a data haven to store your questionable content, look no further than Pionen, which benefits from Sweden's strong freedom of speech and information laws. Past and current clients include the North American Man/Boy Love Association and PRQ, a company that hosted The Pirate Bay and WikiLeaks.

Pionen, White Mountain data center in downtown Stockholm, Sweden. Screengrab: Dean Nelson/YouTube

As the story goes, WikiLeaks was kicked off of Amazon's data servers in 2010 and began storing its content at Pionen. The famous WikiLeaks releases came from two servers at the facility, which the company chairman later auctioned off to benefit Reporters Without Borders, but not before taking a series of brooding photographs with them.

2. In Glasgow, I read in a small data center that was run like an episode of The Office. One of the staffers claimed that the industry was fairly humdrum, so when she and her co-workers get bored, they pull pranks or make cakes that look like data centers. Case in point.

So an amusing thing happened in the run-up to my event: An employee had pitched an article about I'm that angel to Data Centre Solutions, which, for those not in the know, is like the Artforum of the data center industry. The first time I visited the center, we did a photo shoot for the piece.

This may be the single best piece of evidence that I once stood in a data center—this totally suspect picture, where the CEO and I look like cardboard cutouts Photoshopped into a stock image.

Maybe it's a metaphor for the unrepresentability of big data, the incommensurability of the bodily and the informational. I'll let you decide.

3. One last example: I arrived in Haarlem, The Netherlands to perform in a data center. Upon meeting me, the press liaison said that she had read my book. She remarked, "You must be very smart or very stupid, and I have no idea which." This remains the best thing that's been said about it.

The network is far too complex, in short, for the struggle to only play out in physical space.

Our tour guide was the sort of person who flies to conferences on cloud security and knows his way around the discourse. At the end of the tour, one of my guests asked what the future of the internet looks like. Here's what I recall:

He cited a 2012 UN conference on telecommunications governance, attended by representatives of most of the world's countries. The debate revealed a sharp division of opinion in the role a nation should play, with America and the European Union supporting internet freedoms, and China, Russia, Saudi Arabia and others preferring strong—even authoritarian—limits on internet access.

In the wake of the Snowden revelations, we know that this division isn't so sharp. For example, the United States's "super-jurisdiction," as Metahaven claims, extends well beyond the measures afforded by the USA Patriot Act, allowing the government to search and seize data within its territory, as well as from any cloud service that conducts business with the country, no matter its physical location.

In response, we've since seen a partitioning of the cloud: of France and Canada taking precautions against governmental data being routed through U.S. servers—of the rise of securitized private clouds in countries like Switzerland and Luxembourg, which protect your data at considerable cost.

In some cases, this partitioning can play to the advantage of free internet advocates, as with recent policy efforts in Brazil and The Philippines, as well as for countries like Iceland, which Metahaven champions for turning its legal sovereignty, real-world isolation and global connectedness into the foundation of a new experiment with internet democracy.

Nonetheless, our tour guide was fairly pessimistic, seeing the future as one in which governments enforce greater network control in the name of protecting civil liberties—perhaps going so far as soliciting cyber-terrorist attacks to justify increasing security.

Cyber-terrorism can take many forms. You can cut an undersea cable or set off a bomb in the American Northwest, but the greatest threats to the internet, he told us, come through the lines. The network is far too complex, in short, for the struggle to only play out in physical space.


Ever since that tour, I've been turning over his words, because they raise a question that's also pertinent to my project: When no single data center can embody or stand in for the whole, what can we do with this site? Why visit this site?

Even this many years into the project, I wouldn't claim to have answers. In the spirit of James Lee Byars, I'm just looking for questions, and the experience of bringing audience members and data center employees together is each and every time a question: new, different and enduring. Recognizing a data center's incapacity to embody or stand in for the whole—being given access to a transparent view, yet leaving with the feeling that things are more opaque than you ever imagined—discussing energy, security, politics and subjectivity within the unceasing hum of the physical internet: these are not nothings, not just grist for the algorithmic mill.

In general, I've found that these data center events serve as reminders of what Mark Andrejevic calls the "knowledge asymmetry of the big data era": the divide between those who generate data and those who instrumentalize it, as well as the infrastructural divide "shaped by ownership and control of the material resources for data storage and mining." While my events don't structurally change this asymmetry, they are motivated by a belief that we have a claim on these data centers, much as the centers—as the custodians of the internet—already have a claim on us.

Tyler Coburn is an artist and writer based in New York.