Can an Algorithm Solve for Good Taste?

“Culture People Worrying About Culture”—that is, those who continuously refresh Twitter to be part of the discourse—have been worrying about taste (what is and isn’t “proper” to consume or do) quite possibly forever.

More recently, the impacts of internet virality and social media clout on taste cycles have been discussed ad nauseam. See The Cut’s divisive “vibe shift” article last month, which attempts to explain the general confusion in defining the next trend; the “great irony collapse” in Gawker in November about the increasing difficulty of parsing irony from taste; and (Man)Repeller’s article on how the Instagram algorithm has impacted style in 2018.

But the truth of these exegeses remains: personal taste has been flattened by internet algorithms. Prior to the birth of the algorithm, taste was something individuals built up in exhaustive hunts for record store curios. Now, taste is conveniently meted out based on Posts You’ve Liked.

Algorithms work by using artificial intelligence to display relevant content, sorting posts and targeting ads by what it thinks you want to see based on previous engagement: likes, comments, saves and shares. These elements are weighted differently on social platforms, and the inner workings of a platform’s algorithm has become something of a closely guarded secret.

How the internet—and the algorithms driving attention and purchases there—affects taste has given rise to the vaguely pleasing aesthetics driving the success of “blands” and hyper-specific consumption content on apps like TikTok that brought us phenomena like #TikTokMadeMeBuyIt.

Let’s take furniture as a case study. Furniture is a good barometer of taste—in a way that fashion may not be anymore—because of its cost. You buy furniture not just for lasting utility but also for others to see, whether you’re hosting at home or posting on social. And when furniture needs to be enduring, tasteful, Instagram post-worthy and functional, how do you choose what to buy? It’s a problem only an algorithm could solve.

Some criticisms of algorithms emphasize their tendency to automate the status quo by repeating dominant patterns or develop a leaning toward universal palatability in sentiment or aesthetics. The success of homeware blands’ vague and unobtrusive anti-marketing tactics on social algorithms is a good example. Their promise is acceptable, functional, “no one could say it’s ugly” basics. Allbirds but couches, if you will.

There’s another aspect to algorithmic flattening: more equal access to “good” taste. As Gawker’s “irony collapse” piece explains:

“In recent years, the internet-fueled monoculture has both spread and contracted simultaneously, with increasing speed. All but the most obscure items are made accessible with a simple search… When information becomes free and universally accessible, voluminous research for a novel is devalued along with it.’ A similar thing holds for taste: when almost any record or T-shirt or piece of Danish cookware can be found on eBay, are they still impressive to own?”

Looking at it this way, the “flattening” of taste is felt most by the aforementioned “culture people”—at least superficially. Access to good taste traditionally hinges on proximity to insider-y individuals who are hot, young, smart or sceney, and for those eager for taste cred but lacking aesthetic ambitions of their own, the key ingredient: disposable income.

These established taste barriers have virtually vanished. Now it depends on a mid-century modern TikTok reaching a user with enough interest and income for a knockoff Eames Chair or a purchase from a not-yet-burned-out vintage reseller. Or maybe they buy a Kartell x Zara collab T-shirt or get a Le Corbusier tattoo. Same difference.

On the flip side, algorithms use combinations of user behavior, data and AI intuition to serve users hyperspecific content. The TikTok algorithm, for better or worse, seems to know its users well. And the tastes of the niche subcultures it cultivates are at times illegible—or actively distasteful—to outsiders. Look at the recent spectrum of TikTok furniture crazes: cardboard bedframes, bland sleeper sofas, #cluttercore, the infamous “floor people.”

Online discourse often arises around who is included and isolated in the latest taste or trend controversy: are you cheugy, are sigmas cool, is blonde uncool, who survived the vibe shift? But now we’ve been through a few waves of viral taste discussions, there’s a few interesting takeaways.

Even within a hyper-niche community, “good taste” is being supplanted almost as rapidly as trend cycles. (It’s now a bit cheugy to say cheugy, for instance.) And we’re starting to see mainstream brands and artists cultivate niche, ironic personas of their own, tapping into the left-field tastes typical of “meme brands” in an attempt to deliver their corporate brand message with a side of internet-native catnip.

So what’s the outcome of algorithmically determined tastes? Will we all end up in a serene yet characterless metaverse version of The Wing? Or so disparate in our TikTok home decor niches that our friends don’t like our vibes anymore? Hopefully not.

Our taste has long been swayed by contradictory forces: rebellion and conservatism in the ’70s, for instance. Navigating taste cycles on the algorithm probably won’t get easier—for brands or people. But one might find solace in knowing that this tension between mass and niche internet tastes, cultivated by the algorithm, could just be a new version of the time-honored push-pull in culture.

Maybe we’ll even learn to ride the tide more gracefully by learning to appreciate the taste we co-create—whether it’s independently developed or it’s a result of trusty algorithms on TikTok or Overstock.com.