American Slop and the Empire of Optimization
"Optimization" may usher life-saving AI-assisted medical advancement, but with it comes disturbing AI cartoons aimed at children running endlessly on YouTube Shorts. These developments are of a piece.
A writer commenting on “slop” in 2026 is akin to a comedian opining on Donald Trump in 2016. The absurdity of the moment is self-evident, the punchlines practically write themselves, and people can’t get enough of it. Be it essays on “LLM-speak” or furious opinion pieces about AI winning prizes for fiction writing, we seem to be inundated both by slop and increasingly predictable responses from its detractors. Making sense of our tenuous cultural moment may require stepping back and situating the “slop” phenomenon within a broader context of American production.
Slop is often defined as low-quality digital content that is completely or partly generated by an artificial intelligence tool like Grok or ChatGPT. The tools are numerous, but the function—producing quick, automated output, often to turn a quick buck—is all the same. To understand the prevalence and significance of slop requires an assessment of where it comes from. If we look at slop as an approach to production, as opposed to the specific digital output of some AI tool, we can start to understand it with greater clarity, approach it soberly, curb it, and—as is necessary or possible—remove it.
Slop is what it is because it has been produced by a tool that, by design, attempts to flatten, homogenize and create simulacra of what humans may create. Yet these processes are not unique to AI and LLM tools. They exist in various other forms, and the examples are everywhere. Consider McDonald’s, fast-food’s poster child. It may be likened to AI slop in that it flattens, homogenizes, and produces a simulacrum of wholesome food. Be it fast fashion, fast food, or fast content, the urge to produce quickly, produce more, and produce at a lower cost is as American as apple pie. Perhaps we should have anticipated the prevalence of AI slop.
Put simply, slop is a product of American empire. America is where the technologies that produce slop originate—where the “hard science” happens, as it were. Yet even as the infrastructure of slop is American (for now, at least), its presence and effects are certainly global. In the cultural sphere, the same empire that pioneered formulaic cinema (a type of pre-AI slop) and was responsible for the Disneyfication phenomenon (a type of slop-city) was copied the world over. Is it any surprise that “slop content” has similarly exceeded its American locale into all corners of the globe? These are not minor developments; Disneyfication, for example, represents the soft power of American empire. It is therefore worth considering how slop might be more than an anodyne inconvenience.
If slop was not effective, it would not exist. Coco Krumme’s Optimal Illusions: The False Promise of Optimization may give us some clues as to where slop as a way of producing originates. Though written and published just before the slop explosion, Krumme’s central thesis is that the concept of “optimization”—which has its roots in mathematics and engineering—now plays an outsized role in society writ large. This fact is arguably much more visible now than when she first published her book; our algorithms are full of advice on how to optimize our bodies, our schedules, and our work. Through a case study on optimization in agriculture, Krumme argues that efficiency has been “deified” and that the specialization of food production that allowed forlower prices also made fast food possible. Analogously, cheaper, quicker methods of producing research, writing, and entertainment have made slop not only possible, but allowed it to proliferate. The problem here is not that a particular technology has made certain processes easier or more affordable. It is, rather, that the effects of AI are not isolated to the technology’s undeniably useful processes—they are totalizing. There is no life-saving AI-assisted medical diagnosis without disturbing AI cartoons aimed at children running endlessly on YouTube Shorts; these are of a piece.
An interesting distinction Krumme makes between optimization and capitalism is that the former is “top down” while the latter is “bottom up.” With optimization, we start off with the answer to which all processes, guardrails, and variables must combine to produce a singular optimized output. Contrastingly, in capitalism, the market is allowed to “decide” the allocation of resources such that the “answer” arrives at the end of the process. The parallels here to an AI prompt are almost perfect and prescient. Based on the constraints built into a prompt, an AI tool can give you a plethora of results—we start with the answer before the path is traced.
Krumme says that “the wrangling of the world into mathematical models required a set of conceptual shifts that settled into our mathematics over the past centuries.”1 The first of these shifts pertains to the atomization of reality, in which the world becomes something “divided, conquered, measured, and mined.”2 It is hard to ignore the Enlightenment roots of this way of seeing the world. Be it in a novel or a motion picture, many of us increasingly approach a great story less as the messy product of a flawed, unique, and visionary mind, but more as the sum of tiny parts for which we can optimize. Each scene transition—indeed, each word—can be optimized by AI. If the prompt we start with functions to “optimize”—to entertain, to win a literary prize, or to amass ticket sales—then the art itself is secondary. The entire culture becomes beholden to the prompts being written by those with the existing resources to distribute the outputs. As Krumme puts it, “optimization is agnostic to its context. Whether we want productivity or emotional connection, sweet lemonade or sour, the mathematics is the same.”3
The bifurcation of part and whole driven by the culture of optimization makes us lose sight of the hazy mystery of artistic production. In Krumme’s words, it casts aside “things that weren’t readily divisible.”4 So much of slop sits in this gray mean of easily divisible, digestible matter—ingredients just real enough to fend off starvation, but by no means a nourishing diet. Are we forever trapped in this false purgatory?
Given how increasingly health-conscious the public has seemed to become over the last two decades, and with the prevalence of food-critical documentaries like Supersize Me and Cowspiracy, one might be forgiven for believing that the end of fast food was near. In reality, McDonald’s still stands tall as among the most recognizable American brand names globally, and fast food remains a lucrative industry in most major economies. As public sentiment coalesced around the value and importance of healthy food, the market quickly responded with renewed advertising campaigns and met the pressure of documentaries like Cowspiracy, for example, by introducing an explosion of new dairy alternative products. But market adjustments arising from people reacting negatively to the ills of slop are arguably harder to achieve, in part because AI functions to metabolize and utilize all dissent; it will use our well-argued, angry writing about slop to produce refined slop that limits our understanding at best, or manipulates it at worst.
Limiting our consumption of slop is easier said than done, but perhaps one way in which the excesses of our AI-induced culture may be resisted is through the unglamorous halls of policymaking. Producers of fast food are in a seemingly endless tussle with domestic and international regulators regarding additives and ingredients. The pressure these legislative bodies put on producers to ensure better health outcomes, and the results that pressure creates, can often be taken for granted. Public discussions on ultra-processed food, sugar consumption, or specific additives find their way into policy or policy-adjacent arenas the world over. In this way, rigorous research, academic work, and public forums that expose the ills of AI can ensure some level of pushback against the gargantuan tech companies who make slop possible in the first place.
It is likely that we will soon see more research funded by Big Tech that conveniently argues that its own excesses and dangers are overstated, and that these slop-producing technologies possess “scientifically proven” benefits that will usher in an era of unheralded abundance for us all. Academics, activists, and researchers must therefore aim to leverage a well-funded counterattack (or perhaps a preemptive strike) that the main agents of slop may not be prepared for.
Aside from public criticism and struggling with regulators, fast food has also been a victim of the flattening it champions in its own kitchens. The slop cities and towns that produce gray blocks to house fast food restaurants and shops are becoming far less appealing than the colorful restaurants of the 1990s. The soullessness of the retail park mirrors the emptiness of LLM writing filling a gray book sleeve. Even without pushback or policy changes, the blandness of slop itself could be part of its undoing.
It may be that as the human spirit is increasingly tainted by slop, we will come to know that the technologies which produce it are gratuitous at best and ruinous at worst. This may be wishful thinking, and the damage of a numbed audience may already be done, but one can hope that there are enough people who remember a time before the reign of optimized culture took hold.
Coco Krumme, Optimal Illusions: The False Promise of Optimization (New York: Riverhead Books, 2023), 47.
Krumme, Optimal Illusions, 48.
Ibid., 55.
Ibid., 58.
 | A guest post by
|