During a 48-hour-long hackathon event held last weekend, a group of Princeton grad students created What Would I Say? The Facebook app trawls through your content and generates Facebook statuses based on what you’ve already posted. It’s a fun, clickable exercise, but in the days following its creation, it has blown up online. Even for someone like me, who can’t remember the last Facebook status she’s posted since 2008, the process is addicting. When I first started seeing these nonsensical statuses popping up on my news feed a few days ago, I spent the good part of the evening taking screenshot after screenshot of my own randomly generated blurbs.
This is for cockblocking his alone time to knit with a hug/chokehold…
what you find when I don’t know if you actually shot a nice picture.
The thestral is how you go well on fries, but the crossword together.
Trying so pretty.
Bots imitating the way you write are nothing new, and neither is our fascination with them. Earlier this fall, @tofu_product promised to reply with a tweet written like anybody who tweeted at it. That can be my next tweet! follows a similar model, aggregating tweets based on what you’ve previously posted on Twitter. This has all the trappings of a passing Internet fad — fun, shareable, meant to be ephemeral — but there is something more here.
For the most part, the mainstream Internet adheres to strictly prescribed speech patterns. We can joke about the existence of a BuzzFeed story generator, but online content follows certain formats and patterns. Starting a headline with “this” or writing a listicle about “X ______ That Will Make You Feel Like _____” is clickable for whatever psychological reason. Crafting a title that resonates is part of the territory that goes with writing anything, and there’s a bit of comfort for writers and readers in assimilating to a particular Internet speech pattern. But it’s also nice to be surprised sometimes.
Stilted language and errant punctuation feel refreshing, in all their linguistic weirdness. While fed through the voice of spam, these snippets are the type of disjointed poetic content that you’d find scribbled in someone’s notebook rather than a perfectly crafted status update. Twitter feed-cum-art project @Horse_ebooks accumulated hundreds of thousands of followers based on its robotic-sounding content. Nobody’s that nostalgic for spam-like text, but digital nonsense brings a weird creativity that is hilariously only achievable through the automated voice of an algorithm.
This type of generated speech can be more fun than the cadence of “normal” Internet speech. Phrases are jarring and abrupt or long and unwieldy. You’re surprised when a turn of phrase strikes you with fortune cookie wisdom. You laugh at egregious randomized offensiveness. (What Would I Say? doesn’t avoid minefield words like “cancer” if you’ve posted them on your Facebook.) You pause at statuses that recall certain people or places or things, and how they’ve been paired.
Basically, the joy of the app isn’t its ability to predict what you would actually say. It’s in seeing the resulting bricolage of your words, automatically remixed, so different from anything you’d otherwise read in life or online. And that’s something exciting.