A meme is only as good as the number of people who care about it—if no one’s aware of the meme, then it’s not one. Thankfully, the infrastructure that delivers memes to our computer screens increasingly quantifies their significance through hit counts, relevance scores, and follower totals. Without the context those numbers provide, I wouldn’t know whether @Horse_ebooks is my own fever dream, a weird Internet accident that amuses a handful of Twitter users, or a far-reaching phenomenon, LOLed at by tens of thousands. Thanks to the evidence available, I can confirm that it is indeed the latter: At the time of this writing, @Horse_ebooks commands a braggable 62,476 Twitter followers (including myself), with 934 having retweeted its most popular tweet—“Dear reader, You are reading.” Klout, a website that measures influence on Twitter, gives @Horse_ebooks a score of 76 (influential about topics: Horses, eBooks, Knives), higher than famous athlete @lancearmstrong’s score of 73 (influential about Livestrong, Tour de France, Celebrities). Billions of years of evolution have seemingly produced a creature that is optimally suited to thrive in the absurd and fantastical Internet jokescape that the human race finally got around to building.

Literature Review

From what we know, @Horse_ebooks began the way any spambot begins. Someone, supposedly a Russian businessman, created an automated Twitter account to promote the eBooks he sells on his website. The bot copies and tweets random text excerpts with hilariously strange results. The great Jon Hendren “discovered” the account and began retweeting it, before writing possibly the best synopsis of its significance in early 2011:

“@Horse_ebooks is a Twitter bot designed and automated by apparently some Russian guy to sell worthless, horrible ebooks about horses. In order to avoid being detected as a spam bot, it occasionally posts a text snippet or two from one of its ebooks, chosen at random. I will never buy an ebook from it, but I will follow this Twitter account until I die or horses become extinct, whichever comes first.”

On the strength of such early fanaticism, the @Horse_ebooks Twitter account ascended to surprising heights of popularity. Its style, while distinctive, has seemingly changed at certain junctures during the past year, probably due to adjustments of its copy-and-paste algorithm. One such shift last September, documented here, led die-hard followers to conclude that a sentient being had hacked the @Horse_ebooks account, perhaps in order to increase its mainstream appeal (this, of course, was not what had happened). In January, John Herrman penned the definitive @Horse_ebooks analysis (in which the Mountain Goats’ John Darnielle described its tweets as “found poetry”) and, more recently, Gawker’s Adrian Chen tried to find the account’s creator, getting as far as a name (Alexei Kuznetzsov) and location (Tula, a Russian town south of Moscow) but failing to elicit a response from the man himself. One group even applied for a grant to build a statue of the @Horse_ebooks avatar at an upcoming conference.

@Horse_ebooks and the Turing Test

At the center of the @Horse_ebooks phenomenon is a single question: How can a spambot be so funny? Or rather: Why is @Horse_ebooks funny? I can imagine nothing more misguided than picking apart @Horse_ebooks tweets to isolate their humorous elements, but the account’s existence raises fascinating questions about how Internet communication works. The “controversy” surrounding the perceived tone shift of @Horse_ebooks in September 2012 is key: What if someone had hacked @Horse_ebooks? In his Splitsider piece, Herrman quoted a colleague who said that “It’s kind of a signal/noise thing? Like it’s ‘cool’ when it’s us inventing signals out of noise; it’s not cool when it’s all signal masquerading as noise.” In other words, the truly random, as interpreted by you and me, can be hilarious, but disingenuous non-sequiturs written to seem random are much less funny, and at times quite off-putting. If we one day discover that a human being is handcrafting each @Horse_ebooks tweet, however, that person is undoubtedly a genius. Barring that revelation, we must assume @Horse_ebooks tweets are ephemera generated automatically and thoughtlessly by the Internet ecosystem, and like other found art, they are made funny by those who discover, interpret, and retweet them, and not by the @Horse_ebooks bot or even its unwitting creator.

The reason why @Horse_ebooks is so singularly fascinating and so much more significant than a Lolcat or honey badger, or any other meme that remains interesting for 10 seconds, is that it actually passes the Turing test. That experiment, conceived by computer pioneer Alan Turing, inquired whether a machine could communicate with a human so convincingly that a third-party judge would not be able to distinguish between the two. A machine that cannot be told apart from a human by the messages it sends is said to have passed the Turing test. Although @Horse_ebooks is clearly a spambot, it seems just funny enough and just self-aware enough that we can entertain realistic notions of a person writing the account’s tweets, and we can actually speculate that someone once hacked the account to make it weirder and funnier. The fact that I have wondered about this question at least 20 times affirms its near-passage of the Turing test and, to me, explains its soaring popularity: The horse in the avatar actually seems like it’s speaking to you.

@Horse_ebooks may have more in common with natural phenomena like an image of the Virgin Mary in tree sap than with conscious human creations like free verse poetry or Zen koans. Perhaps Alexei Kuznetzsov has successfully crafted a “non-sequitur joke” algorithm. Regardless, the Twitter account has prompted a frenzy of exegesis that rarely accompanies other Internet spam. I’m not joking when I say that @Horse_ebooks helps us to understand what it means to be human. And if you’re not sure whether that statement is, in fact, a joke, that’s exactly how @Horse_ebooks would want it.