This is how Kafka summarizes the post:
The big takeaway here is that while most people on Facebook spend most of their time sharing stuff with a small group of like-minded friends, Facebook is so big — 800 million users! — that Facebook users end up learning lots of stuff from people they barely know: “The information we consume and share on Facebook is actually much more diverse in nature than conventional wisdom might suggest.”
That claim is extremely important to Facebook's business strategy of being able to index its users' online activity and make it available to advertisers so that they can target their ads. (The ACLU has some chilling information about that here.) So naturally Facebook wants to stifle any arguments that using Facebook to do everything would be anything less than enlightening. Kafka urges us to keep in mind that "Facebook thinks you’re getting a whole lot of signal out of that noise" of proliferating updates and "frictionless sharing," but that's almost too generous. Facebook almost certainly knows that information overload is a danger to its prospects and seems desperate here to spin that problem away. The Bakshy post acknowledges that strong ties drive sharing on Facebook, but argues that content from weak ties has the most novelty value, and is more sharable within the cluster of strong ties. "In short, weak ties have the greatest potential to expose their friends to information that they would not have otherwise discovered." This is the basis of Facebook's usefulness as a marketing tool. As Bakshy concludes, in considerably more anodyne language, "online social networks can serve as an important medium for sharing new perspectives, products and world events."
Facebook would thus like to change what we recognize as "signal" and what we experience as "noise" to suit its purposes: getting us to stay logged on and using its apps, and prompting us to authorize more data provision and collection. Signal equals novelty, which equals an opportunity to improve status in the quantitative realm of online social networks. It hopes that we will make more of the random information that comes at us through frictionless sharing and the like go viral within our smaller networks. It would like us to think of sociality and friendship as the potlatch-like exchange of whimsical novelties in a public contest of who can spread more information the furthest. In other words, it hopes to become a more data-rich form of Twitter. Facebook would like us to feel obliged to perform for our close friends on its stage as a way of staying salient to them. It wants our response to information overload to be a fear that we will be forgotten, and that we will respond by distributing more information, worsening the problem for us while improving the metrics for Facebook. Facebook is not an echo chamber but a negative feedback loop.