Welcome to Filterworld

By Kyle Chayka.

12 Jan

Kyle Chayka is a founding editor of Dirt. This essay is drawn from his new book Filterworld: How Algorithms Flattened Culture, which comes out on January 16. Kyle will be on tour through January at bookstores in DC, NYC, Boston, LA, and SF.

TikTok’s For You feed feels like it can read your mind, dredging up interests from your subconscious. Netflix’s homepage tries to guess which movie you’re going to watch next and edits its thumbnails to appeal to you more. Spotify playlists load songs that you’re likely to keep listening to without pausing the infinite stream. Instagram sees you look for a single vintage chair and floods your feed with a hundred accounts shilling the same Eames knockoffs. Google predicts which words you’re going to type in an email, or iMessage in a text message, and offers them up to you as if they are your own thoughts. Everywhere, algorithmic recommendations and tailored feeds are forced into our eyes, ears, and minds. It’s overwhelming, not to mention increasingly boring. 

In my new book, I came up with the word “Filterworld” to describe our interwoven environment of algorithms. These equations have become inescapable, influencing the vast majority of what we consume online — and thus what kinds of culture we consume, period. I use “filter” because algorithmic recommendations are ultimately filters that sort content. They surveil our own actions and the actions of billions of other users on digital platforms, then they use that data to sift out what an individual user is likely to engage with — to listen, watch, hit the like button. What is unlikely to engage becomes invisible. Algorithms also filter in a way that reminds me of Instagram filters: They have a way of warping and distorting culture, remaking art, design, music, or film in the mold of the platform. Hence “Instagram face,” “Spotifycore,” and “TikTok voice.” 

The “world” in Filterworld is because this is the state in which we live, both online and off. Our personal desires are constantly being anticipated, served in advance, and perhaps thus undercut. (Who wants to be faced with the monotony of everything they already know they’re interested in?) Nothing is safe from the influence of algorithmic recommendations. Google Maps routes backed-up highway traffic through a bucolic town unprepared for the cars. Restaurants are forced to tailor their menus more to what looks compelling in TikTok videos than what tastes good. Independent coffee shops all over the world cover their walls with white subway tiles and arrange succulents in small ceramic vessels because that’s what the Instagram audience favors. Everyone buys the same many-pocketed winter coat or the same beige checked rug on Amazon because the marketplace recommends it over and over again. 

The vast interconnection enabled by digital platforms has ended up creating more of a sense of sameness than diversity. Users are subtly guided toward the same subsets of topics, urged on by recommendations that are designed not to serve their interests but to create profitable attention fodder to sell to advertisers. Instagram doesn’t care that you are deeply interested in the history of mid-century modernism, or that you want more recipes for easy weeknight meals. It only matters that you’re still watching and clicking, building a map of your interests that make you more targetable. It’s long past time that we realized these systems are not benefitting culture. Creators and consumers alike are pushed into preset formats that we may ultimately have no organic interest in. By giving in to algorithmic feeds, we are letting tech companies determine our tastes. 

Flatness, like scalability, is efficient. The same culture flows through the same pipes to the same net-average consumer. But since when did efficiency become the sole metric by which we judge art? Filterworld represents the idea that the messiness of culture can be optimized and that only what is optimized for shareability is worthwhile. I disagree.