Normal sounds

Anti-Instagrams

When people need a sense of what success and social inclusion look like, in their immediate circle and in the concentric rings broadening out toward their aspirational horizons, they look at Instagram. It is where users can consume normativity. They can scroll until they have struck some sort of balance between anxiety and reassurance about where they fit into society, and then they can close the app.

Consuming normativity can give us the illusion that we get to sample it when we choose, that it isn’t pressing in on us at all times. But at the same time, the desire to seize control of normativity — to grasp social values in circulation, to recalculate one’s own status and level of conformity, to track the degree to which one has inevitably fallen behind fashion — also generates a certain ambivalence. It confronts us with our lack of autonomy and the instability of our social position; it testifies to the necessity of constant self-presentation (as opposed to being accepted for how our intrinsic being “naturally” appears). Instagram confronts us with our perpetual inadequacy, our dependence on influencers and other kinds of commercial communication to orient ourselves, how all the people we know seem to fall right in line. It’s always there, always refreshing; the feed never stops.

So it is not surprising that there is a continual market for anti-Instagrams: social media platforms that promise to wipe away (or at least permit a temporary escape from) all the social dynamics Instagram trades in — as though Instagram itself were responsible for creating them or making them newly binding — and allowing us to “be authentic” or “be real.” If only we could start again and share images on a different platform, we wouldn’t have to coerce each other into playing out society’s demands; we could appear as we really are, outside all socioeconomic pressures — as though those pressures were in no way formative of our character and our trajectory, as if we aspired to nothing but the freedom to spontaneously display ourselves in our pure singularity. We would all be genuine with each other in some higher realm beyond comparison, competition, and calculation. No castes, no exploitation, no schemes, scams, or strategies. It would be universal friendship.

The New York Times recently reported on one of the newer anti-Instagrams, something called Lapse, which promotes itself with the utopian tagline “friends not followers.” Like the earlier anti-Instagram app, Dispo, Lapse imposes an arbitrary delay between when users take images and when they can see and post them (hence the name). This differentiates it from the much discussed and now fading BeReal, whose gimmick was ambushing users at some undisclosed, arbitrary time with a prompt to share a photo taken that instant.

Arbitrariness always initially seems like an escape from normativity. If you are going to “be in the moment” without any purpose or intention, any stakes or social consequences, as these kinds of apps promise, then that moment has to be arbitrary. Arbitrariness stands in as the opposite of purposiveness. So when users consent to some app-imposed restriction or some narrowly conceived protocol, the very pointlessness of it helps them believe they are reasserting their autonomy. They can feel as though they gain agency by specifically choosing an arbitrary reason to abandon it; they can claim “I am being authentic” when they choose personally to obey rules that don’t seem aligned in any particular way with social demands. At the same time, the gimmick licenses a feeling of acting spontaneously, impulsively, and structures such behavior as “true” rather than inconsiderate or rash.

The gimmick of an app like Lapse (or BeReal, or Dispo, or Daylyy) lets users feel as though they can unilaterally reject what Instagram now represents: the weight of social expectations and the realities of what sorts of performances are valued. In that moment, one can fantasize that the anti-Instagram will restore the “true self” that has been suppressed by having to adapt to social demands, the self that is untouched by all the various forms of influence and temptations to imitation that are meant to warp and recast it.

But eventually, the pretense of giving voice to this “true self” loses its interest, in part because the pristine, uninfluenced self doesn’t exist. If it did exist, it would have nothing to say: It would contain nothing and aspire to nothing. It’s not as though spontaneous or random images that supposedly capture people being indifferent to the expectations of other people are very compelling to those other people either. What makes a platform compelling to keep using is not a fantasy of solipsistic authenticity but the same social normativity that one also occasionally tries to escape.

As new norms begin to cohere on the new platforms and their protocols, their arbitrary restrictions start to seem like a marginal distraction, and social performance and judgment again appear as central. The gimmick gives way to the network. We are consuming normativity again, our own and everyone else’s, and continue on ambivalently until some other gimmick allows us to momentarily dream of a respite.

None of these anti-Instagram platforms can succeed over the long term because a true escape from something like Instagram would require not only getting off social platforms — any social platforms — but also getting out of society itself.

Spotify’s spurious genres

Part of this “visual essay” by Matt Daniels examines how Spotify invents new genres through data analysis. He quotes Glenn McDonald, Spotify’s genre researcher:

We do recognize the existence of genres that are mainly driven by listening patterns and not necessarily by the artists thinking of themselves as being peers … we can find clusters of listeners who share fondness for the same artists, even when it’s not necessarily obvious that those artists share a style, and they don’t necessarily come from the same place.

The word share seems a loaded choice here for describing the “fondness” for certain music that strangers have in common, unbeknownst to themselves. They don’t really “share” anything with each other, in the sense of deliberately disclosing it; instead Spotify imposes a categorization on them and lumps them together regardless of what role music plays in their particular lives. Likewise, artists who may not even co-exist in time or space and may have no knowledge of one another don’t really “share” a style; they are not mutually influential, they are not working with and against one another to establish the boundaries of a movement or an affiliation with stakes for them. Instead they are pigeonholed with a label by a statistician, and that label then has irrevocable consequences for how their music can be received on a streaming platform.

As this 2020 Paper article describes, the names for these pseudo-genres (“Escape Room,” “Braindance,” “Stomp and Holler” etc.) are arbitrarily assigned by McDonald himself in some cases, though one can imagine an LLM generating the names as needed, reversing the text-to-sound models I wrote about last week. Because the genre names don’t stem from any direct social practice, because they never had to catch on but have been imposed by fiat, they grate on the ears, like when a music critic tries too hard to be compressed and clever. “If I described my music taste as drift phonk, math rock, or bedroom pop,” Daniels writes, “it would sound cringey and pretentious.”

It’s no wonder that the musicians and the listeners tend to take these labels as insults. Who wants to be algorithmically taxonomized against their will and independent of their conscious social practice? Who want to be told what their taste adds up to and what it “really” means about who they belong with and what it means? Who wants to be forcibly subsumed under someone else’s set of concepts? As Daniels points out, there are overtones in this categorization process of how music has been racially compartmentalized, reinforcing broader segregation practices. But the algorithmic parsing of individual behavior into a range of ad hoc, statistically deduced classifications tends to treat people as what Deleuze called “dividuals”: as a set of fluid data points to be massaged externally. One can be assigned to any established social classification as a probability, or one can be targeted as a demographic of one. (John Cheney-Lippold’s paper “A New Algorithmic Identity: Soft Biopolitics and the Modulation of Control” explores some of the ramifications of this.)

I really need to read Nick Seaver’s Computing Taste: Algorithms and the Makers of Music Recommendation, which undoubtedly has a lot to say about ad hoc genre production. But in the meantime I will complain in my uninformed fashion about Spotify’s efforts to negate genre by proliferating them without reference to social practice. Spotify imposes genre affiliations as though people don’t put any thought into what kinds of signals they want to send with their consumption choices (regardless of what they are actually listening to privately).

It’s probably a generational quirk of mine that I tend to associate personality with music genres one chooses to adopt, but have those associations disappeared altogether? It’s as though Spotify wants to eliminate those kinds of social determinants for listening behavior and replace it with something that is fully rationalized and individualistic — you should only listen to what is statistically forecast by your previous behavior in relation to certain sounds as mere sounds, isolated from any kind of social resonance they might have, and you should rely on Spotify to generate the music that fits the pattern. Yes, Spotify will compare your behavior with other people’s and feed you content accordingly, but you shouldn’t take cues from your own interactions with other people. You should let Spotify mediate those interactions at the level of data and pattern matching, so that nothing nonstatistical, nothing that pertains to your social connections that Spotify doesn’t control, enters the equation. Who you are, Spotify’s taxonomizing methods suggest, isn’t about social relations at all. It is instead posited as something isolated and self-contained, like the empty “authentic” identity that the anti-Instagrams expect us to want to embrace. Why let other people influence you when a machine can do it all?

I don’t use Spotify, and I don’t really get why Spotify users allow the company to profile their taste in an attempt to automate it. It seems like the price one is forced to pay to use the service. I appreciate that people like to consume content about themselves, but I still find the accounts of Spotify’s “year in listening” reports totally upsetting, as if one’s music consumption was something to track like fuel efficiency. I imagine feeding a bunch of book titles into an LLM and then letting the model tell me what I should have experienced in reading them. Why would I want to save myself the trouble of having my own taste and making my own effort to appreciate things? Why would I want algorithms to like things for me, as if the pleasure wasn’t in the experience but in the list of liked things transmitted after the fact?

Empty fridge

Here is a heavy-handed metaphor for the impact of automation on culture industries: The fully machine-generated AI Seinfeld show, which garnered a lot of attention a few months back, is now stuck in a loop. According to this 404 Media article,

For the last five days or so, one of the main characters of the AI-generated Seinfeld show has been endlessly walking directly into a closed refrigerator. Nothing, Forever, is very broken, stuck on a short, repeating loop for days. It’s also more popular than it’s been in months. 

Maybe that popularity is a hopeful sign: people enjoy seeing AI fail. Or maybe it just means that lots of us want to see what the “boot stamping on a human face—forever” looks like for our times.