Some backstory

I got sick of twitter in 2023, and I decided to hop off; I remember seeing someone on my feed post about moving to cohost, so I decided to give it a go—that’s about it, really. I wish I had some long, elaborate backstory.

This blog post, my longest one in a while, is partially me relaying my experience on cohost to other users, and partially explaining some bits of cohost lore to outsiders who never got on the site. This post will not be a well thought-out history of cohost, a summary of arguments between its users, or an anthropological study of the site—though I will talk about cohost’s culture later in this post—instead, these are just some reflections, some riffs.

This will be disorganized, and this will ramble.

What’s a website to do?

Cohost’s closure was, in my opinion, inevitable. The site was run by four people who refused any kind of voluntary work, split between development, moderation, and maintenance, along with loads of enterprise-level infrastructure, and, most importantly, salaries. In short, the staff were fighting a losing battle as soon as they started—social media is unprofitable, and betting on users subscribing to a premium service is an absolute gamble when your free features are perfectly fine.

Of course, cohost staff’s whole ethos—well, the part about them not accepting voluntary labor—was partially, in my opinion, bullshit. A huge portion of cohost’s ecosystem relied on userscripts for, around a year ago, dark themes, among a bunch of other things. Just recently there’s cohost-dl, which lets users download all of their stuff, including liked posts, off of cohost, effectively doing staff’s work for them. Likewise, one user did all of the work in compiling an exhaustive report on racism on cohost.

There was some great stuff on cohost, though, notably the markdown support and the audio support (which was awesome).

While I don’t want to spend this entire blog post disparaging staff, I think we should criticize the people who built cohost and hold them accountable for their shit, but in a way that isn’t toxic—cohost users could never find the middle ground between coddling grown, adult programmers and treating them like total shit for no reason (and doing Kingdom Hearts villain monologues in their comment sections).

A website built around isolation

funny how a site designed around the comfort and safety of isolation has brought everyone together for its goodbye. it’s funny this site was ever a thing at all

—Rivvy

Users were in bubbles, and it seemed that I would see the same few faces on my feed if I never went out of my way to look at new tags. In some ways, I enjoyed it, I enjoyed not having to constantly see stupid people post stupid things, and I saw posts about things I either enjoyed or found interesting. Cohost organized itself a bit like Reddit, but without official communities—instead, users, including me, picked the places they liked, pulled up a chair, and rarely interacted outside of their little zones.

Posts (or “chosts”¹ if you like whimsy), along with the accounts that made them, did not have readily accessible metrics, meaning that a particular account or post’s popularity was measured not with a like or repost counter, but by the amount of times that a user would see that post or account reposted onto their feed, and the number of comments beneath that post signified both the popularity and controversiality or excitement of that post. There was a distinct vibe on the site, for lack of a better word, that permeated almost every interaction, a sort of “onlineness” that acted as the bedrock for interuser conduct—users were worried about “hijacking” other users’ posts, as though you could get sued for providing extraneous commentary on someone else’s post.

Python code for generating the above chart.

import numpy as np
import matplotlib.pyplot as plt

# Data is taken from the cohost staff's financial reports.
financial_updates = ["August 2022", "September 2022",
                     "October 2022", "November 2022",
                     "December 2022", "H1 2023", "June 2023",
                     "November 2023", "March 2024", "April 2024",
                     "May 2024", "June 2024", "July 2024",
                     "August 2024"]

monthly_active_users = np.array([4314, 3837,
                                 9e3, 37e3,
                                 21e3, 12e3, 22.4e3,
                                 21142, 29846, 18973,
                                 19127, 17868, 18612,
                                 16846])

reverse = slice(None, None, -1)
financial_updates = financial_updates[reverse]
monthly_active_users = monthly_active_users[reverse]


plt.barh(financial_updates, monthly_active_users)
plt.xlabel("monthly active users")
plt.ylabel("financial update")
plt.title("cohost's monthly active users per financial update")
plt.tight_layout()
plt.savefig("2024-10-02_mau.svg")
plt.close()

Cohost’s userbase increased about four-fold after Elon Musk bought twitter, from approximately 9,000 monthly active users to around 37,000. In the wake of a sort of crisis for social media users—where a site they loved (or hated) was about to be run into the ground—cohost presented a new opportunity, a sort of “fresh start” for social media; cohost billed itself as being queer, a place largely immune from the censorship of other social media platforms, and devoid of the pernicious algorithms plaguing other websites. Most of the people I knew on cohost were former Twitter users, and so they carried aspects of Twitter’s culture over.

Put all together—the largely post-Twitter userbase, the isolation, and the fear of trampling on other users’ posts—you get a sort of digital suburb, dominated by cliques of different users, cul-de-sacs. Cohost was thus largely inert, keeping to itself. The number of monthly active users stabilized in June of 2023, though seemed to trend inexorably downward as time went on—the reasons for this are beyond me, but it seems that the number of active users was destined to fizzle out; without any new, large-scale disasters on other social media platforms, cohost was just a slightly less glitchy, less social version of Tumblr.

Maybe I’m being unduly harsh on cohost, because for many people the experience was fair to middling, or even great. Some people found cohost to be a welcoming new home on the web, a place where “people were not sharpening their claws,” as one user—who I can’t remember—put it. Still, I don’t think that means cohost or its sort of culture are above or immune to criticism. As I briefly mentioned in the last section, cohost was predominantly white and largely hostile to nonwhite users, driving many of them off the site. Some users who expressed concerns about cohost’s accessibility were, likewise, also driven from the site due to harassment from other users. Whether or not cohost was a better site than, say, Twitter, is irrelevant—by its own ideals of nondiscovery, cohost created an insular culture which did not take kindly to criticism about the site, and was largely incapable of finding constructive ways to deal with internal disagreements.

Perhaps the cruelest irony of cohost’s culture was the sort of vague left-wing politics held by most of its users—stressing the importance of systemic critique and change—only to suggest that you just download a userscript every time some new site-wide issue popped up. Even if cohost was financially sustainable, its culture was not—users fought passive-aggressively with each other, and there was no way to create a Twitter-like attention economy.

That isn’t to say that cohost was entirely awful. There were many wonderful people on the site who I loved to talk to—I loved reading what they wrote, hearing what they played, seeing what they drew. Perhaps the biggest fault in my own analysis is that I tend to be negative all the time, I’m a born hater, and it’s far easier for me to remember all of the shit than all of cohost’s good parts. More than anything else, there was great potential in cohost, and sometimes cohost lived up to that potential—artists could post all kinds of stuff without risk of censorship, there was an awesome demoscene on the site, and users got creative.

Observations about the site

I first started writing this post on 2024-08-15, less than a month before cohost staff announced their plans to shut down the website. Perhaps it was naïvité, but I left that post in the drafts for a while, like most things—I never expected to log on one day and see that cohost was shutting down after less than five years of operation.

eYada?

In some ways, cohost mirrors eYada, an internet talk-show website that was designed to basically let hosts say anything that they wanted,² getting around FCC regulations and the like. In that same way, though certainly not with the same spirit, cohost was designed, partially, for artists to get around pornography bans on other social media sites—mostly Tumblr, though Twitter is likely not too far behind—by giving them lax rules.

Funilly enough, eYada lasted far longer than cohost, five years as opposed to two. Time will tell if this will mean anything, if cohost will have any long-term effects on internet culture as a whole, or if it will fade into distant, dead obscurity like eYada, a footnote in internet history.

The Website League

There were also a lot of posts saying “cohost may be dying, but it proves what’s possible,” and while I do agree somewhat with that sentiment, I believe it also goes to show the limits of what a website under capitalism can accomplish. The Website League looks to be interesting—and their whole thing about federation, being inspired by social ecology and all that, is a genuinely neat idea, even if it smells faintly like Yesterweb³—though given how it seems to be inhabited almost, if not completely exclusively by popular cohost veterans, I’m not sure if The League will be able to overcome cohost’s problems, as they can’t be magically willed away by citing Bookchin and Öcalan.

There are some things about the League that I think are admirable, such as their seemingly open decision making process, and its federated nature. Indeed, I think that federating the League is actually a pretty good idea, since it should give the League far more longevity, enabling moderation and interaction on the scale of small sites rather than entire centralized networks.

If the League is to survive or, better yet, become successful, it will need to actively work not only to right cohost’s wrongs, but also build a better network of websites. Moreover, the League’s stewards—who, depending on the network’s future trajectory could end up as either a bunch of pals unconcerned about bringing in new users, genuine visionaries, or digital landlords—will need to be accountable and admit fault if, and when something goes awry, and discourage the same kinds of attitudes towards criticism that was engendered on cohost.

There were some good shitposts

I’m not going to say that cohost had better shitposts than other sites—such Tumblr, because cohost never had its own Pjackk—but there were some great jokes on the site, with some particular favorites being

• love honk – Love honk started as a person describing a genre of popular music that a user (I can’t remember who) heard in their dream. In its most traditional form, love honk starts off as cheesy 80s ballad, but vocals are replaced with a continuously honking car horn, though the genre moved beyond that. Love honk was one of the few internal shitposts to break containment, but it wound up never being that popular.
• he’s only carbon now –

  oh geez, just heard the news out of vancouver. ryan reynolds was torched to death by a beam of light from the heavens. yeah the report described him as being “only carbon now” which strikes me as a little insensitive but i guess that’s the modern world for you.

  —Willow (@willow)

• new Garfield – It is what it says on the tin—new Garfield.

Cohost’s humor wasn’t the peak of comedy, but it was refreshingly goofy and earnest.

CSS crimes were so cool

I’ll miss the CSS crimes, where people would use HTML5 and CSS to build all kinds of neat, wacky experiences—games, animations, and pictures that interacted with your page. Honestly, I think CSS crimes were easily the best part about cohost; they were always creative—you can’t get them anywhere else right now, and I hope that some kind of CSS crime-based demoscene pops up somewhere.

Well, now what?

Now that cohost is finished, I have only one major social media site left to browse, Tumblr—and I’m not giving my URL to you. When Tumblr shits the bed, my days on social media will be effectively over—I’ll stick to my blog, my RSS feed, and discord (or whatever cannibalizes it).

Honestly, I think that an RSS feed is a far better method for social media, since it encourages people to actually read and write stuff—with a proper RSS feed, and a decent network of people, you could actually make a fun, interesting, decentralized site of people all working on stuff without the constraints of social media. Some social media services will need to continue existing—because running a blog from scratch can be technically demanding—but that could still be subsumed, somewhat, into RSS. I’m not deluded enough to think that would replace traditional social media—I don’t think you could convince your regular gregarious Greg or Greggette to run a blog and RSS feed—but it would be a network worth being a part of.

Closing thoughts

I think I’ll miss cohost a bit, because there were some features that I really liked, but honestly—if it wasn’t for the cool mutuals I had on here, along with the CSS crimes—I would have stopped using this site a while ago. Cohost’s lack of discoverability and lack of people with my particular interests—mostly physics, synths, and stuff like that—coupled with a sort of general white-queer-tech-Seattleite monoculture made this website quite boring at times.⁴

The Website League is interesting, and I think it’s a worthy experiment, with the goal of creating digital communes. To me now, it seems like now is a perfect time to build new forms of communication, new forms of online existence, but I will likely continue to be both skeptical and cautiously optimistic about The League.

In any case, I’ve archived my cohost page. Thanks to the work of the team behind cohost-dl, I’ll have a tarball of the whole thing that I can keep indefinitely, so that way I’m able to keep all the good parts of cohost. I don’t know if I’m mourning or not, but I see what cohost could have been.

I wonder if this post will be used by some intrepid internet archaeologist in the future—you know, someone doing something like Down the Rabbit Hole, or some kind of lost media analysis. What will they think of cohost? Being a part of something like cohost, an obscure internet project, was interesting, but I wonder if it would even be worth doing something like Cohost again—I’ve become far more skeptical of social media in the past few years and, outside of tumblr, I’ve become largely disconnected from it as well.

I will likely not be searching actively for a new home on social media, since I already have a blog here on neocities and a discord account—I’ll be more than comfortable chatting with some people and writing on here, without the pressure for metrics that other social media sites have. Thinking about it now, I’m wondering if social media is a generally worthwhile project? I would love the current social media landscape to become something more social, more open, more free, and easier—the rise of social media reflects a move towards simplicity for the average internet user and mounting technical costs for the backend.

This post took me far too long to write—in fact, I started writing it in August. This was going to be an incredibly ambitious post—as most things I write on here tend to be—but I cut back the scope quite a bit, since I figured other people would be better at talking about cohost’s history at length.

I’m wondering what to write about next. There are a few things I have in mind:

• Need for Speed Underground 2. I’ve been replaying this game recently, and I have some incredibly mixed feelings. Kacey already did a great review of U2, so I would likely be talking about some points she didn’t really go into, such as the game’s tuning mechanics, its illogical reputation system, or how to get it running on modern PCs.
• Physics. I’ve noticed that every time I started writing about physics on here, I get way too big for my britches and I eventually just avoid writing entirely—and yet, I still want to work on the many-body physics series (which I’ve only written one part of). I have a bit of a roadmap—talking about perturbation theory, classic many-body systems, and modern methods in the field (such as dynamical mean-field theory and the renormalization group)—but I haven’t actually started writing at all.
• Math. I like math quite a bit, but I’m pretty terrible at communicating it to people—I have a tendency to want to start using sets and funky math symbols, so maybe a guide on actually reading math (something which I’m not an expert on) would be interesting.