Cohost and the "caricature problem"

or: an unintentional Cohost requiem

Alyaza Birze (January 12)

The process of enclosure, of carrying out our activities within these private [corporate tech] platforms, changes us, including how we relate to one another and the underlying purpose of those relations. [...] Once physically and legally enclosed, the soil began to be treated as a machine, whose role was to be as productive as possible. So, too, with our online activities, where our relationships and conversations are our modern-day yields, designed to harvest ever more data. As with corn and soy grown in great monocrops, quality and individuality are sacrificed in favor of standardization and homogenization, even when homogenization takes the form of individuals all competing to stand out as quirky and utterly unique.
—Doppelganger, chapter 2

i'm in the middle of reading Naomi Klein's book Doppelganger, and in the second chapter of the book she makes a point about corporate social media (and corporate technology more generally) and how it has enclosed the internet. far from the libertine, do-what-you-want spirit that once characterized the internet—or even the promise of freedom of expression and mass social connection that was effectively the selling point of early social media—we now live in an almost totally enclosed, enshittified internet experience. our social relations online are almost fully privatized and monetized by megacorporations, and the resultant data we inevitably give these corporations is trivially bought, sold, and analyzed by anybody who wants to do that. it is quite bad, and leads to many subsequent bad consequences.

one consequence is, as Klein observes, a mass homogenization—something that likely has an already existing name but which i personally call the caricature problem. i'd like to take some time to sketch this problem out, how Cohost avoided this problem and why it was so enjoyable, and what we'll be losing now that it's gone.

the caricature problem

in short, the caricature problem is this: the internet is so enclosed and so algorithmically captured that it punishes expressing the totalities and complexities of ourselves. to stand out we must exaggerate or flatten ourselves—become caricatures of who we are—into digestible, predictable personalities for consumption. we must in effect become marketable brands. as Klein notes, this is not incongruent with individual quirkiness or uniqueness: actually, those qualities frequently become a manifestation of this problem. even when we would prefer not to, we frequently become pigeonholed into acting out the qualities which other people perceive as defining us. gimmick or single-issue accounts "breaking character" and receiving backlash can be thought of as an example of this—but virtually all of us are subject to this pressure at some point or another, and in one way or another. often times we (preemptively) reduce ourselves like this without even realizing it.

how does this problem come about?

as described, this probably feels like an inevitability. of course social media run by Mark Zuckerberg, Jack Dorsey, Elon Musk, and a cadre of other cretins with psychotic technolibertarian beliefs would render us all caricatures in this way. these are people with a very specific, very inhuman idea of what social relations should look like. but the problem is created by something more fundamental: that we mediate nearly all of our online social relations through private actors at all. the enclosure of the internet provides corporations with such a wealth of data that it is trivial for them—both individually and at scale—to manipulate every aspect of how people interact and in what ways they do so. and of course they have every market incentive to do so (to say nothing of other—frequently political—incentives).

the most prominent and omnipresent manifestation of this is algorithms, without question: how you experience social media today is overwhelmingly determined by mechanical processes that are hidden or obfuscated from you as much as possible. a good social media algorithm is invisible, and yet in its essence reflects the amount of data you (and everyone else) put into it. needless to say, we put an unfathomable amount of data into these private actors—the things we post, who we talk to, the things we like and interact with, and so on—and all but hand them the capacity to "objectify and quantify social life in numerical form," as Richard Seymour puts it in The Twittering Machine. but it should be no surprise that to "quantify social life" in this way produces perverse incentives—rounded personalities are punished, caricatures are rewarded; substantive exchanges with complexities are buried, simple narratives encouraged; and personalities become hostages to their algorithmically-served-up audiences.

this is bad, but perhaps it would be easier to stomach if the ability to "quantify social life" was not simultaneously being sold back to users in the form of metrics, which are actually even more omnipresent. social media is simply awash in numbers: likes, dislikes, reshares, comments, quoted posts, views, follower counts, post numbers, and a frankly endless number of other meaningless units. and this is to say nothing of dashboards that track profile statistics, link clickthroughs, and other business-oriented number values that have little meaning but are usually accessible to literally anyone.¹ it is simply trivial to become a caricature because you are given the ability—in painstaking detail—to know exactly what people want from you, exactly what they don't, and what might make people care or be more social with you the next time you post. this is obviously a recipe for disaster. unless you're explicitly looking for it, who sincerely wants to be on the wrong side of an issue, or the bad end of a ratio, or become the main character of the day and have thousands of people gawking at you for sport? and conversely who doesn't like being reassured they're correct, or having 10,000 people listen to what they say (even if it means compromising on their personality)?

many other factors besides algorithms and metrics inform all of this, of course, but they alone demonstrate the point sufficiently in my view. (they also have unique relevance to Cohost and why the website largely avoided these issues.)

Cohost and its lack of the caricature problem

now, i cannot give you any sense of what should be done about this at the systemic level besides "fundamentally overthrow the existing political, social, and economic system we live in, since replication of this issue is good for business"—and obviously that's a bit of a non-answer. but i do think there are alternatives to this that can exist even within a capitalist framework. Cohost demonstrates this quite well.

ask virtually any Cohost user and they will tell you that Cohost was an incredibly freeing experience for them—a space where they felt the ability to be themselves instead of a living commodity. frequently, users of the site talk about being "deprogrammed" and "detoxed" from the corrosive influence of Twitter and Mastodon metrics, which had previously governed how and what they posted about; others talk about how Cohost finally allowed their art to coexist alongside their other interests. Nicky Flowers writes of the site that "i didn't have to be Nicky Flowers™️, Internet Person Trying To Appear Professional Enough To Hire. no number-go-up, no clout. i was simply nicky." and still another common feeling is that Cohost had a sort of egalitarianism among posters where, because you could only infer influence, even "big accounts" were approachable and able to be socialized with on equal terms.

all of these experiences—and others like them—were clearly and in large part facilitated by Cohost's explicit design choices. from the beginning Cohost committed to no algorithms (instead opting for a chronological timeline and a Tumblr-style tagging system) and a near total absence of metrics (the only numbers was a private-facing notification number, which could be turned off). these were highly-touted selling points for the website, they were substantial talking points in word-of-mouth, and they were extremely well-liked choices by people who used the website actively.² even among skeptics the initial shock and discomfort from the metric choice in particular consistently tended to wear off with time. many of the theoretical problems one might conjure from their absence also simply did not happen in practice.³ "lack of engagement" for example was a reason many artists on Twitter were skeptical of the site's lack of metrics. far from this dampening visible enthusiasm for creative projects though, Cohost had a bustling art scene—especially among writers and musicians, whose mediums of choice mesh poorly with contemporary social media. although anecdotal, it also seems well established that Cohost users were quite generous and had an unusually high click-through rate for artist projects.

i will refrain from drawing strong, sweeping conclusions here due to self-selection bias, but i would argue that Cohost demonstrates quite compellingly that metrics—and algorithms—are completely unnecessary components of a full social media experience. we do not have to be turned into caricatures. it is also entirely possible—even with a small team—to build a functional, modern, well-liked social media platform without either of these. this of course begs the question of why companies don't do this more.

what we will lose with Cohost

i don't need to tell you that the answer is capital. it always is. whether they acknowledge it or not, websites make an explicit, ideological choice to encourage this phenomenon. as Richard Seymour says, data is "one of the most profitable raw materials yet discovered," and yet it is simultaneously easy to induce users to make more of it. you can see a sort of vicious cycle: data is produced by engagement so you juice engagement with algorithms that reflect user inputs. you provide metrics that quantify a user's social life back to them so they know what algorithms and people following them want. those users produce more data. rinse and repeat, the users are—in the most dehumanizing sense possible—the product. Cohost's model eschewed this, so it's an obvious nonstarter for most people in the social media business today. for most social media companies an explicit part of the path to profitability includes selling out their users (usually in multiple ways!) to venture capital firms.

but this illustrates what we will lose without Cohost: an actually existing example of a better world that is possible; a sort of socialism-in-one-website if you want to be tongue-in-cheek about it. it's not as if Cohost ever threatened the existing capitalist hegemony in social media or anything, but you might still analogize its loss to, say, Cuba, or Rojava, or the Zapatistas suddenly collapsing tomorrow. Cohost was—more or less—a genuinely socialist, cooperatively-owned social media company, a beacon of light in an appalling industry. it really sucks to see a radical experiment like this fail, and it's not obvious that another experiment like it will occur given what needed to align to make Cohost happen in the first place. the market is also a harsh mistress: that Cohost lived its values; demonstrated social media might be able to exist without making users a product to be bought and sold; and allowed its users to be real, fully fledged people means devastatingly little. as far as most people are concerned what Cohost stood for is not a viable product because it failed, and the discussion ends there.

but i don't want to say everything is completely dour, even though the probability of another Cohost-like website is obviously low. strictly speaking many Fediverse communities are also examples of this model, and so someone will continue to keep the lights on (just at a much, much smaller scale) where markets almost assuredly won't. in my view the model is also clearly sound, even at Cohost's scale, with better planning and better budgeting. if someone ever scrounges up the money and coordination to make it real, i don't see anything actually stopping a Cohost-like website in principle. maybe a "successful Cohost" in the future will succeed because it learns from its predecessor—becoming the 1917 to Cohost's 1905. i don't know.

if nothing else, though, i hope that the principles of Cohost (there is value in being in the same place as everyone; metrics are ruining our lives; human curation is good and social media algorithms have been a disaster; etc.) continue to inform the immediate future of radical social media. i would like to look back upon this as a "dress rehearsal" to a tech future that does not completely suck ass

footnotes

¹ at one point—and perhaps still, although i have not checked—Twitter provided every user with a dashboard of all of their posts, the amount of interactions they received, the best performing tweet over a given timeframe, and dozens of other measures from which it would be trivial to determine what you should and should not post.

² by far the most consistent critics of either of these choices were users who did not even register for the site, at least in my experience.

³ a lengthy aside i'd just like to put on record somewhere: to the extent that there were actual points of friction introduced by these design choices, these tended to be more inconveniences than legitimate issues with continued usage of Cohost. should tagging or following cool people that interact with your posts have been such a load-bearing aspect of discovery on Cohost? perhaps not, but i would not describe this as an actual issue—this feels too weighty for what was more an inconvenience than anything. many of the issues were closer to future-proofing than anything else. deprioritizing metrics can, for example, be a legitimate impediment to identifying where abuse is actually coming from. on Cohost this could have eventually become a serious issue since it was possible to sort of "quote retweet" a post, and none of these notified the person whose post was being "quote retweeted". but to my knowledge this was never a substantial vector for harassment; certainly Cohost was never lacking in harassment or harassers who could have done this. in any case, i think i would rather grapple with the complexity of squaring that circle than put up with Twitter metrics because it's slightly easier to trace who is imploring their 100,000 followers to harass you.