aA aA
abc abc
Losing Yourself in the Dark
Caroline Busta
Jak Ritger, illustration of interconnected sub-clearnet communities spilling into each other, sharing flows of content, 2021

If one were to imagine a spatial layout of the internet—not the cables and satellites and server farms (albeit fascinating to map in their own right), but the metaverse—what does one see? On screen, professional emails and personal DMs coexist with public tweets, pirated movies and notarized tax documents, each separated by millimeters. We accept this. Our devices are just the user-level portals to the digital beyond. But what does this beyond look like? What are its regions? How does one travel between these spaces psychologically or, moreover, psycho-geographically?

In the past year, our analogue neighborhoods came to feel like the backstage to everything that was happening “virtually” (if such a distinction still has any meaning). With all “non-essential” activity turning online-only, bodies outside the screen were newly free to look and do as they pleased. Dress twenty-something and gothy one day, middle-aged professional the next. Out for a jog? Pass the teens synchronized-dancing in 15-second bursts. Observe their indifference as an onlooker irony-posts this performance to her own gram. The teens don’t care because their primary audience will never see this passerby’s content. And although they’re neighbors in the analogue world, the teens and the passerby will likely never knowingly meet online, physical proximity being incidental to spaces of digital belonging. Both parties have become cyber-local-prime, their “regional” communities algorithmically organized into groupings of similar market-profiles regardless of the geographical distribution of their members.

So how do we map this, and how do we understand the communities that form in each space? What affordances and pressures define the various digital neighborhoods and what kinds of environments do they produce? Creators featured in KW’s Open Secret are no doubt aware of these questions, encountering them as they carve out enclaves beyond Web 2.0. By this, I mean digital spaces that are not indexed by mainstream search engines and that do not strictly correlate users with their government-registered selves. This sub-clearnet (i.e., non-Facebook, Instagram, Twitter) strata of the internet is growing rapidly, changing the shape of the social web in turn.

Michael K. Bergman, “Harvesting the Deep and Surface Web
with a Directed Query Engine,” 2001

To start we might consider how, for the past 15 years, the internet that most users know has been built around the individual as the smallest divisible unit—the user as pixel (contrast this to the pre-millennial television era where, despite attempts to further differentiate it, the household was functionally the base particle). Since the mid-00s, “going online” has meant logging on and being automatically fed a personalized world: your personalized Google search, your personalized news feed, your profile page, your follower count, your private messages, new posts tagging you. And what could be more engrossing than a digital mirror? TV is full of other people. But on the internet, you are the center of the universe—as if in a reverse Copernican turn. Meanwhile, smart devices with their casual “i”s (styled lowercase, like the tech-overlord presenting to shareholders in a t-shirt) have not only reinforced the illusion of one’s personal gravitational pull, they’ve successfully individuated the capacity for productivity and attention of each user so that they can be hyper-efficiently utilized (not to say exploited) by third party commercial entities. In this way, Margaret Thatcher’s famous creed—“Who is society? There is no such thing! There are individual men and women”—was supplied with its perfect tech.

But it turns out that when 53% of the world’s population fragments into 4.2 billion i-pixels,{1} we find ourselves much more vulnerable to top-down extraction, peer-to-peer competition, and the illusion of agency through swarm-like formations that, given this infrastructure, ultimately serve the goals of the platforms more than those driving the swarm. However, it also turns out that this is not the only internet.

As a way of imagining other internets, let’s first give the one that is presently hegemonic two more dimensions:

If one key parameter for mapping online space has been how the online space isolates and regroups its users (the individual-user-as-pixel isolated from its physical surrounds placed within a new digital local according to shared consumer profiles), the surface web / deep web schema has been another. Proposed in 2000 by information scientist Michael K. Bergman, the “deep web” model of the internet entered more popular use in the following years with the rise of illicit digital marketplaces and the broader adoption of Tor, Hushmail, and other tools for encrypted web communication.{2} At some point, Bergman’s initial illustration of a fishing trawler sitting atop the ocean gave way to the now well-circulated (and thoroughly memed and détourned) “iceberg” diagram of the internet: above the water’s surface, the peppy sans serif logos of mass platforms in primary colors in full sunlight; immersed just below, WikiLeaks, 4chan, and dark marketplaces such as PirateBay and SilkRoad, before the berg descends into more sordid and depraved forms of encrypted internet use. In the popular imagination, this y-axis—spanning Instagram-your-brunch to file-sharing-snuff—illustrates the user-pixel’s possible range of motion.

Deep Web infographic typical of web marketing and or online security services circa 2018.

A third basic parameter is the geopolitically specific digital “stack” one must use to access the surface web (or clearnet). Users in China, for instance, do so via the BAT stack (Baidu, Alibaba, Tencent). In the US (and by extension, other NATO member-states and official Global Partners), the clearnet runs on GAFA (Google, Apple, Facebook, Amazon). In Russia and many post-Soviet regions, mail.ru (Vkontakte, Odnoklassniki) and Yandex define this mainstream online space. Meanwhile, residents of the global south connect via a patchwork of these aforementioned platforms as each vies for market dominance. Whatever the physical location of the entity supplying your internet connection (unless using a virtual private network to circumvent your local ISP), your clearnet activity—social media use, web searches, map services, shopping, etc—will be automatically bound by the digital stack that governs that region. It suffices to say that this, in turn, very efficiently shapes the way each region’s users perceive the world.

From these three vectors, we can start to imagine some kind of spatialization of the mainstream internet: the user as pixel, active along a y-axis running from the clearnet to deep web spaces, accessing them via a particular geopolitical digital stack (encryption workarounds notwithstanding).

But this model is insufficient for understanding our place within the broader (non-”you”) internet and how we are as various “we”s online—communities that may share multiple enclaves along the vertical iceberg axis. It also perpetuates old conceptions of how online space is organized, which limits our ability to autonomously (non-algorithmically) find content—and each other.

New Models, schematic of the Dark Forest as habitable sub-Clearnet zone, 2020

Recently, something called the “Dead Internet” theory has emerged from 4chan’s /x/ board. Its premise is that the indexable (i.e., clearnet/surface) web is a “bloated corpse,” a sprawling, listless entity swollen with auto-generated content and fake profiles. Proponents of this concept note that “compared to ten years ago, the internet feels … empty.”{3} Personally speaking, I think it does too—which should be unnerving given that, ten years ago, it contained half as many (2.1 billion less) users.{4} But what if the clearnet platforms are just organizing us differently or filtering who we see more intensely than they were 10 years ago (siloing us into increasingly homogenous groups); and what if, simultaneously, we’re simply not active online in the same way we were then? What if, as social media has come to resemble stretches of strip malls anchored by mega-churches, users are spending more time in the sub-clearnet zone, beneath the water’s surface or—to use a different metaphor—in the dark forest,{5} where one can forage for content semi-anonymously rather than having it forcefed algorithmically. What if the issue isn’t with the internet being un/dead but with our working model of it being outmoded?

I’m not totally sure what this new model would look like—and I recognize that there is some benefit to keeping things obscure (and thus less easy to recuperate)—but here are some notes that might be useful in sketching out alternate forms:

Individuals are not at the center of the internet; industry is.

Human users account for only a fraction of global web use, and our clearnet access is made possible not because it’s some inalienable human right but because our presence there is itself a commodity. Framed this way, it’s interesting to think about how social media data runs through the same fiber optic cables as data governing oil refineries, waste management systems, power grids, and shipping logs (among other industrial IoT tech), as well as massive amounts of data generated by financial markets, scientific research centers, governments, and so on. Representing the social web as just one component of a spectrum of global connectivity might help us think outside of the hegemonic idea of the internet as you-centered and also demystify the Earth-scale systems that are both monitoring and changing the planet.

The clearnet is a b2b ecosystem not conducive to individual human life

In recent years, clearnet platforms have nudged users toward a personal-professional status (blue checkmark certification, prosumer “creator dashboards”)—which honestly makes sense—given the degree to which users are scrutinized (by both platform bots and other humans) and held accountable for what they say (online), not to mention the expectation that users self-financialize, leveraging their private lives for platform success. Rather than lamenting this, we could instead simply acknowledge that social media is not the personal, local space  that it presents itself as, but an explicitly professional, airport-like non-place for performing the “minimum viable spectacle” (Peter Limberg, founder of the online community The Stoa) necessary for reminding the world wide web that you exist (and are available to work). On this speculative new map, we could imagine social media represented more like a grand bazaar, with lanes of kiosks, grouped roughly by trade, displaying representative works to passersby. At the back of the mini-shop is a trap door with stairs leading to a sub-basement where deals can be done. This sub-basement is connected to other sub-basements and ultimately reaches out to the edge of the medina where a neutral commons allows workers, children, and older people to interface directly, unmediated by the protocols of the bazaar.

Drawing of Constantinople, 1521

The smallest divisible particle may be the individual user, but “individuals” can also be run by groups.

This is the essence of incorporation. And if the clearnet is asking us to function as corporate entities, there is no reason, especially in the era of the lowercase “I,” that incorporation can’t be casual, too. Finstas do this already, with multiple anonymous users posting to a shared account. In MMO, gaming success is shared and depends on support from one’s guild. With the rise in dark forest creator communities such as New Models and Joshua Citarella’s Super Secret Sleeper Cell, users have the option to avoid being personally indexed on the clearnet and can instead air their ideas as part of a collective voice. This engenders more experimental thinking because the consequences of a bad take are buffered by the collective shell and the benefits of a viral take are shared rather than building the singular celebrity of (and thereby further isolating) any individual member. This is an interesting exploit of the hegemonic map of the internet, as it calls for ‘overclocking’ the “1 ‘i’ = 1 individual” formula that underpins clearnet economics.

We need a neutral commons

Without light that leaks in from the mainstream, dark forest communities can become myopic and atrophy. There needs to be a selectively-permeable barrier between the clearnet and the dark forest, but also a place for various dark forest communities to communicate with each other. Web 2.0 social media is currently the default commons. Yet there is little tolerance on mainstream platforms for transgressive expression and tripping the censors can get you indefinitely shadowbanned (i.e. in the dark, without light) or worse. This is true of both the GAFA stack and the BAT stack despite their ideological opposition. One simple solution for this is to embrace DIY printing; for instance, one could periodically precipitate a physical object from a community’s online activity and allow that object to circulate through IRL social networks. But it would also be cool if guilds of dark forest communities could come together to build independent digital commons as well. Perhaps the affordances of Web3—where community stakeholders can collectively direct an enclave’s financial resources toward a shared goal—will help facilitate this in the near future.

In practice, the new map will be continuously changing, waxing and waning in complexity as the internet continues to evolve. For now, though, the most important thing may first of all be to make the maps messier, undo the closed loops, venture away from the you-centered internet and get lost amid the more unhinged data flows—both on- and offline.

Broad-stoke impression of geopolitical digital stack regions, 2021


{1} As of January 2021, global internet penetration reached 59.5 percent, with 53% of the world’s 7.9 billion people active on social media. Per: https://www.statista.com/statistics/617136/digital-population-worldwide

{2} See: MK Bergman, “The Deep Web: Surfacing Hidden Value,” Taking License (Aug. 2001) https://doi.org/10.3998/3336451.0007.104. The iceberg diagram first gained traction on Web 1.0 messageboard sites like bodybuilding.com, particularly with the rise of darknet marketplaces.

{3} See also: Pseudiom’s video Dead Internet Theory, The Internet is Empty, April 2021 https://www.youtube.com/watch?v=BEIZHlfjAT8 and @IlluminatiPirate, “Dead Internet Theory: Most of the Internet is Fake,” Agora Road forum, January 5, 2021 https://forum.agoraroad.com/index.php?threads/dead-internet-theory-most-of-the-internet-is-fake.3011/

{4} https://www.pingdom.com/blog/internet-2011-in-numbers/

{5} “Dark forest” is a term coined by Yancey Strickler in 2019 after Liu Cixin’s 2008 SciFi novel by this name. https://onezero.medium.com/the-dark-forest-theory-of-the-internet-7dc3e68a7cb1