The AutoContent Wizard, Part 2

Herodotus, a historian who makes up for his frequent lack of accuracy with his fondness for a good yarn, tells a strange story about Smerdis, a king who once ruled some of the far eastern provinces of the First Persian Empire. During a political intrigue, Smerdis was assassinated, but information in those days didn’t travel particularly quickly, so none of his subjects knew that he had died. As a result, his place was taken by an enterprising magician, also known in some sources as Smerdis, who ruled the kingdom for seven months, pretending to be the old king, before being killed himself in a palace coup. I thought of this story—which I first encountered in a passing reference to “the imposter magician Smerdis” in Jorge Luis Borges’s “Tlön, Uqbar, Orbis Tertius”—while I was writing yesterday’s blog post about PowerPoint, since it seemed like a good analogy for our current situation. We’re told repeatedly that content is king, but the king’s place has really been usurped by our own imposter magician: the AutoContent Wizard and his descendants. It’s the kind of “content” that renders all ideas equal, as long as they fit within the existing template, and it’s no accident that such sites tend to be built by entrepreneurs whose success or failure hinges on their ability to use PowerPoint to convince investors to give them money. A deck isn’t just a tool; it’s part of the foundational myth of these companies. And the content that we get from this mentality is more or less exactly what you’d expect.

Take Emerson Spartz, the tech entrepreneur who created the Harry Potter site MuggleNet, but who has since thrown in his loyalty to a different sort of wizard. Spartz, who has since founded such viral content generators as Memestache and GivesMeHope, is a master of the PowerPoint deck. Earlier this year, a New Yorker profile by Andrew Marantz, which still makes my blood boil whenever I read it, described a talk that he gave at the Millennial Impact Conference, the first slide of which read: “Hi! I’m Emerson Spartz. I want to change the world.” The presentation included the kind of short, pithy tips, legible in thirty-point type, of which such decks are made: “Try to change every comma to a period.” “Use lists whenever possible. Lists just hijack the brain’s neural circuitry.” And Spartz seems serenely convinced that there’s no subject so complicated that it can’t be shoehorned into a viral format:

If I were running a more hard-news-oriented media company and I wanted to inform people about Uganda, first, I would look it up and find out exactly what’s going on there. Then I would find a few really poignant images or story lines, ones that create a lot of resonant emotion, and I would make those into a short video—under three minutes—with clear, simple words and statistics. Short, declarative sentences. And at the end I’d give people something they can do, something to feel hopeful about.

And while we might think that such sites are simply giving readers what they want, there’s a critical point to be made here, which is where the AutoContent Wizard analogy really shines: like PowerPoint, formats like the slideshow and listicle aren’t made for the enjoyment of the audience, but for the convenience of the creator. Content mills that churn out dozens of articles every day, most consisting of material appropriated from other news sites with minimal attribution, need a few basic templates into which stories can be repackaged without conscious effort. A string of images and captions joined by a loose theme, like celebrity hairstyles, is about as mindless a format as could be imagined, and a writer can crank out ten of these in the time it would take to write a single well-considered article. Content generation is a numbers game: if one piece out in a thousand goes viral, you throw as much as possible at the wall to see what sticks, and the advantage lies with models that can ramp up the number in the denominator. The fact that the result is relatively painless to read is a fortunate side effect, but it’s incidental, or one of many factors under consideration. Like fast food, it’s calibrated just as much for cost, convenience, and volume as it is for the fact that it goes down easy. And while I enjoy a Big Mac as much as anyone, there’s no question that it takes up too much of our diet these days.

So how do we depose our imposter king? The first step is to acknowledge the AutoContent Wizard for what he is: an invention of content producers, who have an enormous incentive to convince us, the consumers, that this is what we like. The second step, I’d argue, is to make finer distinctions of quality between the kind of content we read—not just between listicles and longreads, which anyone can do, but between good listicles and bad ones. (As a rule of thumb, if an article just repackages content from elsewhere, it’s probably not worth your attention: like a copy of a copy, the details are degraded each time it gets reproduced.) There are funny, artful listicles and slideshows, and there’s no shame in using them to kill time, as long as you aren’t also killing journalism in the process. The listicle model isn’t going away, so the least we can do is draw a line between those that use the format to entertain, divert, and even inform us and those that utilize the template to generate empty clickbait. This may seem like a tenuous point, but Pauline Kael’s reminder holds as true here as it does anywhere else: “Movies are so rarely great art that if we cannot appreciate great trash we have very little reason to be interested in them.” Or, as Robinson Davies put it: “God save us from reading nothing but the best.” Not every article has to change lives, but the AutoContent Wizard is gambling precisely on our inability to tell the difference between real entertainment and its simulation. And we shouldn’t be shy about pointing fingers. As Lucy Lawless once said on The Simpsons: “Whenever you notice something like that, a wizard did it.”