Another four miles

I don’t listen to podcasts. Perhaps it is because I don’t have enough time killing to do — I don’t have a commute or an immiserating job — or perhaps it’s because I lack the curiosity about the lives of specific other people necessary to listen to protracted, aimless interviews. I may be too awkward and curmudgeonly by nature to find any appeal in parasociality, let alone the ordinary kind of sociality. It’s not unusual for me, when compelled to attend a party, to bring a crossword puzzle along to serve as both a pacifier and a shield; brandishing it allows me to immediately beat a retreat into my fortress of introversion.

The genre of podcast where people “hang out” and engage in unstructured chitchat for hours about a basket of loosely related topics is especially alien to me; it sounds like drive-time radio for a journey that everyone has accepted will never end, because there is nowhere to go and no way to get there. Livestreams presumably take that banter-and-blather ethos to its full phatic apotheosis, where, as Eric Harvey suggested in a Bluesky post, it becomes “just recursive content, a person or a bunch of people sitting in front of a camera talking about how many people are currently watching them sit in front of a camera.” It is just presence announcing itself as the only topic left, the only thing atomized monads can have in common.

Harvey is responding in passing to a promotional link to this Taylor Lorenz newsletter, whose headline commands, “You Should Be Paying Attention to Kai Cenat.” Do I have to? Really? Apparently he has received 32.2 millions of hours of metered attention on a streaming platform, so statistically speaking, every person on earth has given him a few minutes of attention already. Lorenz claims that such numbers indicate that “live streaming is becoming more pervasive and a bigger part of our media diets. People are craving interactive events and content, and traditional media struggles to replicate the participatory experience that live streamers provide.” I am obviously not one of these “people” with this craving (the parade of proper names of internet creators in Lorenz’s posts are even more foreign to me than most of the names in the Coachella flyer), so I could clearly benefit from some insight into why that is. Simply pointing to the leaderboard doesn’t explain anything.

This contrast between the interactive content that “people crave” and the “traditional media” that has become outmoded has taken on a political cast since the election. In a New Yorker item, Joshua Rothman argues that “Democrats preach while Republicans riff” — i.e. Democrats operate like the nattering nabobs of traditional media while Republicans have mastered a less inhibited approach to providing the vibes and parasociality that people crave more than factual information. He posits that a new kind of “exciting and surprising” politician — a.k.a. Trump — thrives amid the media overproduction that technology has wrought:

it’s evident that politicians, too, can thrive by producing vast quantities of malleable, interesting content. A politician can post and tweet and retweet; he can speak off the cuff in memorable ways, encouraging people to film and upload his remarks. He can go on podcasts or give epic, rambling speeches, talking for hours with all sorts of people, so that clips of the best bits can be shared. He can cultivate multiple media channels—not just staid press conferences and interviews but also social-media platforms, merchandise, and memes. If he does all this, he can become dominant and inescapable, rising above the sea of information while his opponents sink beneath the waves.

Rothman likens this space of devalued verbiage “in which words are produced constantly, and hardly matter” to a “process of collective improvisation” between politicians and audiences that “internalizes the dynamics of the Internet, transforming them into a political stance,” though that sounds a lot like the livestream that talks about nothing but the livestream. In the ambiance of simulated co-presence, all the isolates project their own terms of recognition onto the leader and pretend that they are somehow self-fulfilling. Meanwhile moribund politicians cling to “messages drafted by institutions” that “don’t rise above the sea of information.” They are tuned out because they have nothing of the livestream in them, no sense that something unpredictable might happen — even if the livestream’s inchoateness is just as predictable in its own way.

In her piece about Trump as a “podcast bro,” Tressie McMillan Cottom notes that “if you tune out their words (and who among us doesn’t tune out when listening to a podcast) you are still consuming the cadence and texture of the podcast bro style.” That texture remains the same regardless of the content or its political valence; McMillan Cottom likens it “reactionarism” and to the consumption of media to express identity, feel vicarious feelings, and indulge in the pseudo-participation that people crave. The podcast or livestream as a form is politically reactionary in that it emphasizes immediate and transient feeling as the horizon of political experience.

But Joe Rogan doesn’t seem to be all that different from Howard Stern or Don Imus or Rush Limbaugh at that level of texture. Talk radio (and especially sports talk radio) was no less parasocial than podcasts, and it consistently leaned toward outrage fodder and reactionary fulmination capable of mediating a partial sustained attention over hours and hours of discourse. That is to say that immersive reactionary media isn’t a recent invention and doesn’t require the specific conditions of media overproduction that currently adhere. It is not a hot new trend or innovative new technology, but a “traditional” mass media building block, going back to the “fascist agitators” Adorno liked to write about, as in this essay.

According to Adorno, the agitators’ “speeches themselves are so monotonous that one meets with endless repetitions as soon as one is acquainted with the very limited number of stock devices. As a matter of fact, constant reiteration and scarcity of ideas are indispensable ingredients of the entire technique.” That seems to be the opposite of the “exciting and surprising” politics that Rothman hears in Trump, but they are really describing the same thing: the content is monotonous so that the affective dimension can become central, and feel fresh and engaging. The podcasts are diffuse and rambling so that their emotional current can remain primary and palpable over varying levels of listener attention and engagement. They feel participatory even if, and perhaps especially when, you are not really listening.

If technology is making the illusion of participation feel more potent, it’s not by overproducing the means of human connection but by creating the conditions in which endless, aimless discourse has no phatic component. Rothman begins his piece with a strange anecdote in which he reports having a conversation with an AI researcher at a cocktail event.

“Have you tried ChatGPT’s Advanced Voice Mode?” he asked me. (I had.) “The conversations you can have with it are almost as good as the median conversation you can have with a person!” We laughed, self-conscious about both our small talk and the contrast implicit in what we were discussing. We’d spent the evening in one linguistic world — a heightened one, in which every word mattered. We were now describing another world, in which words could be produced endlessly, and hardly mattered at all. Technology seemed to be ushering us from the first world into the second.

Perhaps that’s true, but the whole setup is so odd: Who could possibly believe that an interaction with an AI model is (1) a conversation and (2) that it warrants evaluation in those terms, as if it prompting a model were equivalent in anyway to reciprocal interaction with a living being. Who thinks of “median” conversations, as if they are all being rated quantitatively? Who ranks them? The low-information output of podcasts operates outside of such evaluations; they are not efficient delivery systems of useful data.

Being a podcast listener seems to have little in common with chatting with AI model, even if one can use generative tools to make ersatz podcasts. If this analysis is on track, no one would ever listen to an AI podcast for any other reason that novelty, because what people want out of podcasts is the reactionary wave of affect. In listening to a podcast, there is some belief held out for the libidinal value of that kind of sociality even at several mediated removes — a sensitivity to that collective emotional current remains.

Rothman considers this possibility:

talking about A.I. and surrounded by amiable chitchat, I wondered whether, someday soon, conversations with human beings would be deemed lacking if they didn’t exhibit chatbot-like speed and responsiveness. Maybe there are some circles, in tech or elsewhere, where the quality of “the median conversation you can have with a person” is already measured unfavorably against the yardstick of A.I. Or perhaps the opposite is true: maybe we’re coming to value the awkwardness, vulnerability, and spontaneity of human conversation even more.

The very structure of this makes me want to argue that these are not opposites but dialectically linked. Human conversation is valorized on the basis of its seeming strictly nostalgic, no longer within reach or to be expected as an ordinary part of the human experience; AI is valorized as the technology that makes conversation slow and superfluous. Podcasters and livestreamers perform the lost possibility of pointless conversation precisely as something you need to consume and experience vicariously, something fundamentally transactional — a commodified form of the resistance to commodification that once manifested in small talk. At the same time, no one will be expected to remember what “the median conversation you can have with a person” was like or how it was possible that all “conversations” didn’t supply exactly the same sort of gratification, no matter what they were about.