In a Politico Magazine article last month by Calder McHugh, “Charlie Kirk Memes Have Taken Over The Internet,” I was quoted saying:
“The internet wanted to take control of his image,” said Walker. “People took these Kirkification memes and used them as a way to destabilize whatever image [conservatives] wanted to create. And so I don’t think anyone will be able to look at Kirk 100 percent seriously now.”
I stand by that quote, but after saying it on the phone interview I wondered what I actually meant — why does control of an image matter so much? And, as an operation to take control of an image, why did Kirkification matter in its moment? The following post is in two sections, the first a sketch towards a general account of “the image” right now, and the second a more specific account of Kirk’s face as “an image.”
The Image: From Evidence to Engagement
People tend to think of images as superficial, less important than the real-world things they purport to represent. This tendency still seems to hold in political circles that are not MAGA or MAGA-adjacent. Whether it’s centrist Yglesiascels who declare “slow boring” policy questions matter more than flashy proposals, or leftists who ask “what did y’all think decolonization meant? vibes? papers? essays?” there is a default preference for the real thing over its representations. This preference is paradoxically (and predictably) intensified by the fact that it is usually a pose — “the lady doth protest too much, methinks.”
Of course, a preference for the real thing is common sense. People should strive to not be exaggerating, lying, or submerging themselves in fantasy. But that common sense should be nuanced a little further in the contemporary moment because images are not what they used to be.
Images always had value as a means of persuasion, mobilization, and communication — not just as containers for arguments, ideas, and facts, but as particularly graceful and portable containers. They retain that value. But when images are posts first and pictures of things second (as they are today) they take on another kind of value which is more foreign and fickle. Paradoxically, the image today is both more and less material than it used to be, because it is a post first.
A post operates differently than a picture, even when the picture is also broadcasted on a screen. A post is a live document rather than a static slice of time. It is monetizable on the basis of precisely measured live engagement, tends to travel more swiftly and flexibly, bobs along without anchor on an opaque sea of algorithmic buzz. Additionally, the post is a site of exchange, interaction, and engagement. A viral reel, quote-tweeted meme, or photo of the bruise on the President’s hand is not just a representation of a thing, but a sparring-ground for commenters, a hook for the algorithm to bait, and a revenue source for its maker. The post is a volatile electron in the ever-roiling collider of social media platforms as much as it is a representation of the world.
Put otherwise, the post is an address (in both senses of the word) on the internet: a place where things can happen and money can be made, but also a moment at which algorithms and audiences may be called to attention. In this way, a post is more material than pictures on a broadcast, in a book, or in someone’s head are — it is live, present, and imbued with kinetic energy in a way mere images of things are not. But it is also less material because it is worse at representing real experience outside of the internet.
Ruby Justice Thelot writes in his essay “The Death of the Image”:
…when images overtake reality, they stop representing reality; they stop serving as evidence. They change and are no longer guides to make sense of the world. Rather, they exist as independent entities.
Thelot (who btw I was recently in conversation with in NYC for the release of my pamphlet On Skibidi) points out two frightening trends in visual culture: First, the rapid proliferation of AI generated pictures taken out of data rather than real visual experience and the accompanying possibility that there were more images generated by AI than there were taken by cameras last year; and second, the sense of “default fakeness” which accompanies the process of acquiring knowledge about the world through images on the internet today. Unlike in the past, everybody now thinks every image is fake-until-proven-otherwise — they always kind of were, as Susan Sontag taught us — but“pics or it didn’t happen” no longer applies because every day you see pics of things that never happened. My cherished mutual Jeremy Carrasco, an AI expert who forensically analyzes videos that followers send to him to determine if they are real or not, reports that 76% of all the videos people send him are actually real.
So maybe in 2026, the image can no longer “mean” primarily by reference to a thing in the world outside of itself because the delicate social contract which sustained the use of images as natural evidence has unravelled. Now, as a post, an image “means” primarily through the situations it participates in, by the engagement metrics and signals it gives off.
The same thing, it seems, has occurred with language to some degree, and the example below maybe illustrates it clearly. As I argued last week, the most meaningful two words in the text of the viral hype post about Clavicular’s behavior at the club are the clickable “show more” because the text matters more as a post soliciting engagement than as visual evidence of reality, which is simply a field where you farm posts.1
There is need for a larger analysis of the different ways our conventional use of images and language have been deranged in recent years. There are many similarities — each was disrupted by an AI innovation (image generators and LLMs respectively) and each are now funneled through social media platforms that distort them, their makers, their audiences. But there are also differences, both in the nature of visual expression versus linguistic expression, and (maybe) in the way algorithms process them, namely by breaking apart a post’s audio, visual, and metadata components and then weaving them back together.
But tabling that question for now, I want to say that both images (and language) are losing that evidence value and gaining value as engagement.2
In my two previous posts about Clavicular (and more or less throughout this blog), I have argued we’re entering a world where measurement by machines supplants meaning-making as a way of orienting ourselves in the world. Measurement derives from engagement, while meaning derives from evidence. I’m quoting myself here because I described this idea better in that previous post than I can now:
Meaning is dependent on who you ask, but measurement is dependent on what you do, or are. Rather than navigating weird, tangled, and unfair systems of human meaning-making, people like Clavicular choose to optimize for metrics. Instead of learning to talk to women, you simply do steroids because a person’s attractiveness can’t be a basket of mostly-intangible qualities, it must be a battery of measurable quantities — canthal tilt, facial width-to-height ratio, maxillary recession. And your only logical task in life is to “maxx” those metrics.
This way of thinking naturalizes (and reflects) the dominance of algorithms, which are constantly scanning the social world for stuff they can count and then resculpting that world so they can count it better and push the numbers to where companies would like them to be… the algorithm is not just a tool used by a platform on its users, but an all-around context into which everything — from economic transaction, to personal relationships, to our understanding of current events — is embedded, as essential to being a contemporary person as a body of water is to being a goose.
The goose-in-water-ness of the algorithmic order makes it, to my mind, a form of ideology in an Althusserian way — the thing which “interpellates us as subjects.” Ideology is not “conservative” or “liberal” here, but a technical term denoting the water the goose swims in.3 Ideology does not script our behavior so much as provide us the language we write our own lines in, the stage we perform on, and the standards of legibility which make us recognizable to the audience. Images, when serving as evidence, are valuable because they help us know who we are, who other people are, and what things in the world can be in relation to us. And after the advent of photography, images are a way for us to tell one another what and who is real, while pretending that it is not just us deciding that question of what is real.
But under a newer system — what I call “algorithmic ideology” in the Clav posts — it’s not as much about processes of social distinction, negotiation, and institutionalization which are elaborated by discourse and through evidence. Rather, a lot of the authority is with machines. Algorithms and AI (albeit with human input) decide what is seen and what is relevant. The image-as-post is crucial to this newer system of measurement because it becomes the probe stuck into the social body to determine how things are shaped at a given moment. What mark the image left in the data set, what conversations it has sparked, and what calculations it has changed constitute the main ways that it alters the world/your experience of it.
Two things to add. First, this shift isn’t entirely a rupture: images in the old way always were a little like posts, and posts in the new way can still serve as evidence if people choose to read them like that. Things are blurred. Second, this is neither progress nor decline — it is shitty, but so was the old system; it has its merits, but so did the old system. Both progress and decline narratives about the internet are lazy, unproductive bullshit. For a quarter century, everyone told an uncritical progress story about technology, and now that it seems tech is kind of evil, they’ve just inverted it and are now telling a story of regression, blaming the phones for everything. This is a good way to get views, but it is not a good way to understand the world. The interesting consequential thing is not whether phone is good or bad, but that phone is here and we can decide how to use it.
An Image: Charlie Kirk, from Poster Boy to Posted Brainrot
People are deciding how to use the phone and the networks it gives them access to every day, in real time. Memes are examples of these usages.
I set this whole thing up about a thousand words ago to talk about Kirkification, and after all this stuff about “the image” it’s time to turn to “an image,” by which I mean Charlie Kirk’s face, which was everywhere on the internet in the last few months of 2025.
Kirk was undoubtedly famous before he was killed. But it’s safe to say that the majority of images ever produced of Kirk, by volume, were produced in the last four months. I couldn’t open TikTok from September up until a few weeks ago without seeing Kirk’s face. Many, if not most, of these images were made using AI, were clearly not real, and were never intended by their makers to seem real. In other words, these images had no evidence-value — instead, they had engagement-value. The meaningful thing was the post’s participation in the “Kirkification trend” — and then the hundreds, thousands, often millions of likes, comments, saves, and reposts also participating in the trend through engagement with the post.
In the following weeks, memes moved beyond the core Kirkification trend (which I described in a previous post) and Kirk was shoehorned into other lores, formats, and meme-making traditions. Charlie Kirk is in the Agartha lore for the same reasons the word “skibidi” is in “sticking out your gyatt for the rizzler / you’re so skibidi / you’re so fanum tax.” Each of these elements is a piece taken from a different viral meme familiar to the audience and proven to generate engagement.
But it should be mentioned that meaning is not entirely evacuated, because there is an adjacency in terms of niche: Kirk fits alongside Yakub, the idea of Agartha, and the white pharaoh because he comes from rightwing internet vibes, just as Skibidi fits with gyatt and fanum tax because, like those other terms, it comes from online brainrot vibes. So the logic of combining different viral memes together is not random, but based on niches.
A more direct example (since, like “sticking out your gyatt for the rizzler,” it is word-centric) might be the slangmaxxing trend of “lowkirkenuinely.” Adam Aleksic’s analysis of this is amazing, and I’d follow his lead in calling it microbrainrot. Like many meta-memes, it emerged later than the primary memes made after Kirk’s death, coming about in November. At first, “lowkenuinely” (lowkey + genuinely) was coined, and then “Kirk” was interjected by others — performing verbally to the new slang word what the Kirkifcation trend performed visually to the faces of the people on the internet. The meme trend then followed the crescendo logic that many memes do: lowkirkenuinely (lowkey + Kirk + genuinely) becomes “lowkirkentologixsevenlowstate” (lowkey + Kirk + genuinely + ontological + 67 + flow state). It should be noted that the lengthy, drawn-out explanations of an arcane linguistic code here are presaged by “Onika=Burgers” and other memes of that nature — just as the characteristic appearance of Kirk microbrainrot on TikTok (the poster’s face concealed behind a wall of explanatory text) calls back to other formats of posting.
But the important thing is that the component pieces of the word are more meaningful as elements that have been crammed together in a ridiculous way to demonstrate whimsy and algorithmic overload than as expressions of semantic sense. At this point, engagement value is made to replace evidence value. I see this as an example of a work of art serving to expose, through provocation, a tendency that typically remains underexpressed — this meme trend is about how you can be meaningless as evidence but eminently meaningful as engagement. To steal a phrase that Ruby Justice Thelot used when we were discussing 6-7 during the On Skibidi panel, it’s “playing with the physics of internet, but with no particles — usually you have to adopt some kind of meaning or signified you’re pointing at…”
I love this metaphor and was thinking of it when, about a thousand words ago, I described the post as “a volatile electron in the ever-roiling collider of social media platforms as much as it is a representation of the world.” I don’t know a lot about physics, but I am led to think electrons are a kind of subatomic particle without mass — but which still have certain effects on other particles and the forces that bind or break apart atoms, molecules, and the world. Engagement, then, could be understood as a kind of electrical current — both in the literal sense that it happens through electricity, and in a more figurative sense that exists as a type of static between ideas, words, and images. The power of electricity to generate force was understood in ancient times — people saw lightning, were shocked by static after petting their cats — but it was not really instrumentalized until the Industrial Revolution. A similar shift may be happening with the power of images: evidence-value is like water power, horse power, or combustion, but engagement-value may be like electrical power — a property previously invisible but now emphasized by a new technology.
And so taking control of Kirk’s image through engagement — reversing the current, to perhaps irresponsibly extend my metaphor — might be seen as a new way of articulating power and persuasion. Those who idealize Kirk have not been refuted or silenced so much as they have seen the discursive field shift around them, moved by opaque currents of engagement that alter how and why things may mean. And it is far from the first or last time that has happened to an image.