Hi folks, my apologies for the radio silence in recent weeks, but I’m back to talk about one of my favourite/most hated topics: TikTok.
As Garbage Day’s Ryan Broderick recently wrote, “The number one content trend of 2023, as far as I’m concerned, is TikTok rage bait.” We have seen countless moral panics centre on TikTok, often when a “trend” is shared on another platform, usually first on Twitter where media people still hang out, or otherwise re-posted on Instagram Reels, and no context is carried over. This is called “breaking containment,” as something that might make some amount of sense to one audience is shared with an audience that has no bearing on it at all. Because we live in an unbearable time, these TikTok moral panics often end up becoming conservative bait, particularly in a moment when the platform itself remains the bipartisan target of politicians and other cranks foaming at the mouth for another Cold War with China. The result, as you most likely saw last week, was a fervour over TikTok teens going ham for Osama bin Laden.
If you missed that, don’t worry about it too much — basically, a few dozen users shared videos about bin Laden’s “Letter to America,” written directly after 9/11, and did so without much context; a few of them went slightly viral by TikTok’s standards (more on that in a second); the media picked up on it after The Guardian responded to the interest by taking down their version of the letter; chaos ensued and the Biden administration felt compelled to comment on it.
What that frenzy illustrates, though, is a couple of things about TikTok, and how little people still seem to understand about how it works. First of all, “views” work on TikTok in a more intensified way than perhaps any other platform (other than, maybe, Twitter’s Musk-era view inflation). Unlike YouTube, for instance, which counts a video as “viewed” after 30 seconds or so, TikTok seems to consider a “view” to be counted after less than a second. This essentially means that anyone that simply scrolls past the video is counted as a viewer, which means that the most popular “Letter to America” TikToks, viewed around 500,000 times, are hardly considered to have gone truly viral on that platform, when you really need to hit millions of views to be considered a viral success. For this reason, if you see news reports talking about the mass audience for certain videos and they refer to any number under a million, I wouldn’t take it very seriously.
If anything, this feels like a repeat of how the media covered virality in the early days of social media, misunderstanding the scale and reach of the web at that time, and the meaning behind the numbers. TikTok may challenge the existing standards that were established in that era, but reporters should be smarter about what’s actually happening nevertheless.
A larger issue, though, is documenting the actual influence of TikTok, especially when it does come to ideological content. What does the platform allow, what affordances does it provide, what kind of user gains value? This seems pertinent, too, in light of the similar brouhaha over the supposed pro-Palestine bent of TikTok, and the way Israel’s genocide has been shared, disseminated, and censored on the platform.
In her book The Influencer Industry, which I have almost certainly mentioned before, Emily Hund writes:
authenticity is not just a social construction but an industrial one, continually tussled over by a sophisticated and complicated profit-making enterprise whose decisions about what expressions of reality are valuable help determine what types of content and tools for communication and self-expression are available to the world’s billions of social media users.
While this accurately captures, I’d argue, the overall state of affairs that the social media platforms of this era inculcated, it also seems to subtly misalign cause and effect, or at least it fudges somewhat the directionality of purpose within the industry. What I mean is that she claims that it is the social and industrial construction of authenticity, ostensibly by users/influencers but also their agents, managers, and branding companies, that determined “what types of content and tools” were made available as a result. This has certainly been true in some cases, but I argue that it gets the situation largely backward, as in the vast majority of cases it is the users and influencers who have been made to respond and adapt to the tools and forms of content provided to them by the platforms. This may seem like a trivial distinction to make, as the flows do proceed in both directions in a general sense, but it remains significant nonetheless to point out the innate power differential in platform affordances — users are always made to act within the limits set by the platforms they populate.
It mirrors the argument made in Taylor Lorenz’s new book Extremely Online: The Untold Story of Fame, Influence and Power on the Internet. While I tend to appreciate Lorenz’s tech reporting, the book is framed in a way that argues the platforms responded to the demands of users that sought to find ways to make a living online. I’m confused as to why Lorenz is so quick to align the hyper-commodification of the self in the social media age to the influencers that she profiles, suggesting that this notion of self-branding for profit is an inherent human trait, rather than the result of capitalism, duh, but moreover the structures of the platforms we have been trained on. It’s true, as Lorenz points out, that many of these sites, including Tumblr, Instagram, and TikTok, began without many tools or features for users to make money, shop, or even encounter ads. However, the adjustments they would proceed to make strike me not as concessions made to influencers, but as technical answers to profit-making difficulties in platform capitalism.
Take Twitch. The defining moment for the livestreaming website, in my eyes, is the introduction of “Bits,” which allows viewers to very easily donate money to streamers they admire. They were pitched to streamers as a convenient way for them to monetize their fan community, without having to do anything much on their end at all. In reality, Bits were a response to a profit problem, as up until that point, streamers utilized third-party websites like Patreon, Paypal, or others to facilitate similar payments — money that Twitch saw none of. Now, Twitch earns 30% of every Bits purchase. This is platform enclosure 101, folks, sold to users as convenience but obfuscating the profit motive.
To me, this is how we should understand the history, up to now, of social platforms and the influence of influencers. Many of these companies take their time building a user base, learning from their habits not in an effort to appease what they want, but to introduce ways to monetize them, to extract evermore exchange value, under the guise of benevolent creator-focused agency. What does this have to do with TikTok moral panics? An influencer’s real talent, on TikTok especially, is awareness of what the algorithm may want or reward, and navigating it while simultaneously maintaining a veneer of authenticity. This talent becomes the primary selling point of the influencer as a desirable role, the job that many children claim to aspire to, and it is this ability that, inevitably, likewise renders it within our discursive online experience as not only habitual and aspirational, but non-exploitative and trustworthy.
An illustrative example of what this looks like in practice is the phenomenon of “NPC streams” on TikTok, which garnered much attention earlier this year — another moral panic. Some TikTokers, particularly by using its livestreaming function, began to act like NPCs — background “non-playable characters” in video games that tend to repeat the same non sequitur phrases and perform the same movements over and over, all the while making blank or exaggerated facial expressions like a pixelated virtual character would. This had been happening for some time on TikTok and other platforms, including Twitch, but the practice gained mainstream media attention when one user, the Montreal-based Pinkydoll, had her videos shared all over the internet and go viral among normies. In the videos, she would continuously say things like “Yes, yes, yes… mmm, mmm, ice cream so good, ice cream so good,” to the pleasure of her viewers and the utter bewilderment of those unfamiliar.
Here, as Christine H. Tran argued, not only were we witnessing the logic of the influencer in action, but also “just the latest genre of creators who divide their bodies into marketplaces of intimacy.” As Tran suggests, this is an economy of intimacy (or “intimacy”) whereby creators give over their bodies in exchange for donations, subscriptions, and gifts, participating in algorithmically-successful trends and utilizing the platform’s full suite of tools to keep attention and money flowing.
As Tran explains:
Buying an NPC streamer a virtual gift triggers a chain reaction of the absurd: the gift will appear in a creator’s stream as emojis like ice cream cones, balloons, or even squids. NPC streamers then react accordingly to the specific image on their screen. Squeals, catchphrases, lip smacking and a myriad of gestures result. These broadcasts offer built-in incentives to keep engagement going.
Noting this is not exactly new, as streamers on Twitch, earlier camming websites, and elsewhere have undertaken the same strategies, the NPC streamers nevertheless “render their bodies as intimate interfaces,” and have thereby “gamed our eyeballs into profit” — by gaming the algorithm, if it needs to be said. In addition, NPC streamers like Pinkydoll can also leverage one platform’s success into another, as she encouraged fans to also subscribe to her OnlyFans for more explicitly adult content. “The economics of online fame are closely tied to a more invisible economy of online sex work,” Tran argues, which recognizes, among other things, how gaming the system according to platform trends and demands, and the maintenance of social relations as transactional intimacy, shape us into consumers of the absurd. In that case, looking here for reliable information begins to seem to make as much sense as, well, “ice cream so good, ice cream so good.”
Ephemera
If you happen to care to learn the full story behind what has been going on with OpenAI, you can go through Max Read’s summary — be warned, it’s so stupid, in that it’s not even interesting corporate intrigue, it’s just annoying.
Something light: The Washington Post’s Department of Data went through the numbers on jump scares in horror movies over the decades.
Excellent Rest of World report on “how China became the world’s shopping cart.”
The workers AI hides: “Amazon allows Requesters to set qualifications for their jobs, such as a minimum number of HITs completed, or a minimum acceptance rate. The platform’s default setting is 99 per cent, meaning for workers to be able to access the lion’s share of jobs on the platform they can only afford to face one rejection for every 100 jobs they complete.”
Song Rec: “Your Spit” by IAN SWEET