effect: “July 23, 2 AM: Kate Losse and Samuel Henley are in a complicated relationship.” Oh, I thought, I get it. Sam was testing the new product, News Feed, which would launch weeks later. Engineers and customer support were always testing as part of our preparation for product launches, trying to find the bugs in a feature before launch, and around the office testing was a good occasion for a joke on everyone else. We could get away with anything if we said it was a test.
Still, I was taken aback by the fact that the site had literally written a story about us and distributed it to our friends, illustrated with a photograph that had been posted to Facebook a few weeks earlier, of Sam and me in the pool at the summer house. This fully articulated story, written and illustrated by a machine, meant that authorship was no longer human but algorithmic; we didn’t write our own stories anymore. In the photograph the algorithm chose to illustrate our new relationship (to provide visuals for a relationship story, News Feed finds a photo in which both parties to the new relationship have been tagged), we were dangling over the edge of the summer-house pool, our bodies trailing off into the water, and Sam was smiling straight into the camera, while I look slightly off center, quizzical. Thewater was a beautiful green, pacific, surrounded by darkness. It was an entrancing picture and I could see why everyone, regardless of the fact that Sam and I were just friends, wanted the story to be true. Everyone loves love stories, even if they are just the byproduct of a quality assurance test.
This new product was intended, as all engineering innovations are, to produce efficiency, allowing us to consume content about our friends more easily and automatically than before. However, the publicity provided by News Feed, the way it functioned like a newspaper, did more than just feed information more efficiently. It established a world in which anything that happened to us became food for a narrative, in which we became like characters in a novel that Facebook and its algorithms were writing, whether we wanted to be in it or not.
As was the case with all new features, we had already been using News Feed for months before it launched. As I lay around the pool house with my laptop, watching people play, I was also reading the newspaper-style updates that would appear in my feed. They were usually photo albums from fellow coworkers—pictures of parties at the summer house and elsewhere around Palo Alto.
The general concept of News Feed was simple: An algorithm was now surfacing content that it believed, based on your activity on the site (what you looked at), you would find interesting. But like all technology, the social news generated by a computer lacked some of the nuance of the real-life gossip channels it replicated. Information that would have gotten to you via human contact and conversation now surfaced as impersonally as if you were reading the New York Times (or more aptly, People ).
As News Feed was nearing completion in August 2006, I was sitting on the gray modern couches in a sunny alcove on the engineering floor, testing the feature, when Pasha, the product manager in charge of News Feed and the only woman with an engineering background at Facebook, asked me to review some of the stories for wording. “I’m not good at this,” she said, “You are. Help me.” I supposed that she had turned to me because at that point I had already developed a reputation as the literary one, due to my status updates composed of music and literary quotes and my general disinterest in saying anything absolutely literal.
Pasha handed me a printout of the News Feed stories the team had prepared. At first glance they were, technically, neat: pulling profile photos and updates from the story’s characters to create an algorithmically generated story. However, as I read through them I cringed a bit—they were not about telling a meaningful
Salman Rushdie
Ed Lynskey
Anthony Litton
Herman Cain
Bernhard Schlink
Calista Fox
RJ Astruc
Neil Pasricha
Frankie Robertson
Kathryn Caskie