Though Upworthy has, through its headlines, popularized Loewenstein’s idea of a curiosity gap, he hadn’t heard of the website until March, when I asked him to look at it. I waited on the phone while he navigated to Upworthy’s homepage from his office at the London School of Economics, where he’s on sabbatical. Upworthy’s headlines create logic puzzles, powerfully enticing people to solve by reading. He starts with: “3 Non-White Actresses Show Up to An Audition. I Think We Know Where This is Going.” “The person who wrote this clearly knows we don’t know where this is going,” he says. “But if you read that you wonder: ‘What’s so obvious that I don’t understand?’ ”

Quickly, he dissects the pattern. “A lot of curiosity is due to violated expectations: You become curious when you expect to see something and you see something else.”

But Loewenstein adds that curiosity doesn’t always lead people to respond to a story. “When curiosity is satisfied, the result is generally disappointing,” he said. “The pleasure you get from getting information is incommensurate.”

Applying the curiosity gap to serious news—the kind that comes without an inspirational or anger-inducing message—has in practice proven difficult. Readers were horrified when CNN briefly commandeered the strategy, tweeting “What state has the highest rate of rape in the country? It may surprise you,” and “14-year-old girl stabbed her little sister 40 times, police say. The reason why will shock you.”

It’s tempting to think that these formulas provide a fail-safe way to use the internet to pass along content. And yet even Upworthy admits that nothing is foolproof. “I was talking to students a couple of days ago about the question, ‘How do you make something go viral?’ ” says Critchfield. “And it really struck me that this is not a question for me, this is a question for God. I cannot make something go viral from my sheer determination or will. The audience has to be there.”

And even if Upworthy has found a formula to compel readers to share content, it doesn’t necessarily translate to those readers engaging in action. Some social scientists see reason to believe that supporting a cause publicly can make it less likely a person will participate in subsequent actions—an effect that translates to liking a Facebook page or sharing a Facebook group. “You’re less likely to [give support] now that you’ve satisfied that need to look good to others,” says Kirk Kristofferson, a doctoral student in marketing at the University of British Columbia who recently published a paper looking at the effects of such digital “slacktivism.”

That may be why the most successful efforts to produce true social engagement involve action and sharing in a single activity—an insight commandeered by Mark Zuckerberg when he decided to increase organ donation using Facebook. Clicking a box simultaneously registered a user for the organ donor list and shared the act with friends—triggering a sweep of activity that increased organ donor enrollment in 44 states by a factor of 23, according to MIT Technology Review. Some of Upworthy’s most shared stories have led to action, but on a relatively modest scale, considering the number of views of each video. A video about a group of kids creating a renewable energy campaign attracted 1.6 million views and raised $150,000 for the kids’ campaign. It’s a far larger swath of money than the kids would’ve attracted without Upworthy, but small in comparison to the people that watched, shared, and did nothing.

Even if Upworthy’s posts don’t trigger legions to action, the fact that they can trigger action at all is a powerful antidote to the filter bubble Pariser bemoaned in 2011. What Upworthy has proven is that understanding, and rigorously testing, the kinds of things that drive people to share content can turn that random behavior into something predictable. It’s a formula that can overcome Facebook’s algorithm and, in a way, act as its own algorithm—catering to our most base needs in order to commandeer attention. If Upworthy has figured out a formula for directing attention, they’ve figured out a new kind of formula—one that makes the content they choose to promote increasingly influential and important.

It’s a force that might make Upworthy more powerful as it inches closer to producing its own content. In March Upworthy announced partnerships with news organizations like Climate Nexus and ProPublica. Though the details of the unions are still being ironed out, and will likely just involve cross-promotion of material, Upworthy isn’t ruling out branching into original content. Koechley said they are thinking of themselves as “Netflix”—gathering data and waiting to find holes in the market.

Alexis Sobel Fitts is a senior writer at CJR. Follow her on Twitter at @fittsofalexis.