banality of the algorithm

This Tuesday, my father and I drove up to Bard College to pay our respects to the grave of Hannah Arendt.

At the base of her headstone, nestled between the clusters of visitation stones, a dozen pens and pencils poked out of the ground in homage to her writing.

Looking at the pens, it occurred to me that I never would’ve thought to leave one myself if I hadn’t seen other people do the same thing. In fact, most of those pens likely wouldn’t have been there unless someone had the thought to place theirs first.

Whoever put down that initial pen, then, made it easier for everybody else to also leave their own pens, by introducing a possibility in our minds for us to act on.¹

This phenomenon in the insect world is known as stigmergy: a mechanism of indirect coordination, where each action leaves a “trace” that makes it easier for a succeeding action to occur. The textbook example is how an ant secretes pheromones in its path, creating a shared memory for other ants to follow the same route with reduced effort. As each individual is subsequently rewarded for taking the easier path, they inadvertently make it even easier for others to do the same.

Humans are also very susceptible to stigmergy. When I’m making my way down a crowded sidewalk, I’m always struck by how people naturally sort themselves into fluid columns, following others going in the same direction. It’s easier, of course, than trying to fight someone headed the opposite way—but then we also make it easier for the person behind us to move forward.

Beyond motion, this affects almost all of our social behaviors, especially in how we spread memes and language. It simply takes less mental effort and social risk to follow what others do, so we perpetuate trends—often reaping social benefits that reinforce our behavior. It feels good to place a pen by an author’s grave, so we do that, and then others follow in our example.

This particular trend is an especially apt way to honor Hannah Arendt, whose most famous work focused on the stigmergy of bureaucracy. To her, much of the Holocaust was perpetrated not out of sociopathic malice, but out of a banality of evil—an institutional complacency, resulting from political structures, that made it easier to perpetuate terrible actions. Even many top Nazi officials, at the end of the day, acted not out of monstrous intent, but out of mechanical complicity—ants following pheromones pointing to professional promotion or social prestige.

The Holocaust is an extreme, tragic example, but banality is responsible for so many of our collective actions, because institutions are always the path of least resistance. It’s simply easier to mold ourselves to social structures, in the same way that it’s easier to go with the flow of the crowd instead of against it.

Upon interviewing the creators behind viral racist videos, for example, I was surprised to discover that none of them seemed to be generating those videos out of actual racist intent. Yes, the racism was there, but their primary motivations were more banal factors like “views,” “virality,” and “followers.” Many were also taking cues from other edgy content creators, who were all taking cues from Meta’s recent decision to significantly loosen AI guardrails. In short, these Reels got millions of views because some Instagram executive decided that slop is better for their profit margins, and that eventually trickled downstream into our feeds.

The more I study cultural trends online, the more I see this cycle replicate, even for the most insignificant memes. As soon as a new idea becomes popular, creators hop onto the trend, indirectly coordinating to kill the meme in pursuit of clicks. Since words are memes, the same is true of language, which is why I think our “brainrot” vocabulary emerged from the over-commodification of speech.

I want to stress that stigmergy is not all bad. “Banality of evil” and “banality of the algorithm” are just two examples of how the phenomenon can turn negative, but it can also make life much easier.

Language, for example, is a positive example. If we had to invent new words for each new sentence, we would never understand each other. Instead, we tap into “traces” of how people have used language in the past—indirectly coordinating to communicate through a culturally conditioned understanding of how communication occurs. In that sense, stigmergy is great. It’s just important to know where your ant column is going.

---

If you liked this essay, please consider pre-ordering my book Algospeak, where I directly address many of these themes through the lens of online language change!! Pre-ordering is the best possible way you can support my work, so thank you <3

1. One of my recent videos examined this phenomenon from the perspective of stationary store notebooks: some initial marking always creates a suggestion for how future people should also write in those shared spaces.