10 min read1 hour ago

–

The truth about social media, echo chambers, and how your feed controls you more than you think

Press enter or click to view image in full size

Photo by Gaspar Uhas on Unsplash

Once every few weeks, I go to my Instagram search bar and enter a random word. I peruse the search results and spend time consuming arbitrary posts, exploring new accounts, and watching full videos. On TikTok, Facebook, and YouTube, I’ll replicate the same process, immersing myself entirely in a topic I don’t really care about for 10 minutes at a time.

Call it a desperate bid for agency or a genuine attempt to expand my content horizons; either way, it doesn’t work. Whether it’s stamp collecting, ping-pong, or bird watching, my time spent in the temple of the topic at hand has little-to-no staying power.

The algorithm is a flowing stream. It may slightly redirect when it meets a rock or downed tree branch, or an impromptu binge of pimple-popping content, but ultimately, it always corrects its flow. It knows where it’s really headed, and any attempts to hoodwink it or step outside of its current will be met with placation and swift redirection.

… Or will it?

What Is “The Algorithm”?

The term “algorithm” itself has become something of a boogeyman; a catch-all term for a series of mechanisms and subsystems that contribute to our curated social media landscapes.

Algorithms are, essentially, sets of instructions. They’re mathematically-backed marching orders intended to systemically solve a given problem. In the case of social media, the “problem” in question is “what type of content should the user be served?”.

If we’re being technical, what we refer to as social media “algorithms” aren’t just algorithms — they’re heuristics.

Heuristics resemble algorithms in that they also embody a series of instructions by which a problem may be solved, but, while algorithms exist to determine a precise and objectively correct answer, the objective of a heuristic is to produce a solution that is good enough for the problem at hand.

**Heuristics like the ones that gatekeep our daily content are more concerned with having any answer than they are with what that answer is.

More than any kind of evil mastermind or Skynet product, the algorithm is a perpetual motion machine. What it’s doing is less important than that it keeps doing it.

There is no architect sitting in a room full of screens, secretly giving the algorithm every command. The actual process by which our agency is being hijacked is a lot more banal. The truth is, nobody knows exactly how the algorithm works. And that’s kind of the point.

We’re All In This Together (Kinda)

Part of the reason it isn’t easyto game your algorithm is that it’s not technically *your *algorithm. Not exclusively.

Although we think of social media as the ultimate isolation machine, it is, at the end of the day, a network. It links us closely and permanently — not in a warm “kumbaya” sense, but in a manner that’s hard and calculated.

A big consideration that the algorithm makes is how your first-degree connections are behaving, where you overlap, and where you don’t. That “where you don’t” part is nuanced. Yes, interests that are unique to you and meaningful deviations from the consensus behavior of your community do factor into the content you’re served, but not all anomalies are treated equally.

Spikes in curiosity like the ones I described at the outset of this article don’t tend to stick. What sticks are patterns; large waves of behavior in the larger sea of users.

The reason my attempts to throw the algorithm off don’t work is that my algorithm is basing its decisions on far more than just my actions, and my random searches are barely a drop in the bucket of relevant data.

So, now we know that our individual actions simply aren’t important enough to influence the algorithm alone. This doesn’t bode well for our central question of “beating the algorithm”. But maybe there’s another way?

Making Patterns

Press enter or click to view image in full size

Photo by Swello on Unsplash

“If you get a piece of candy immediately every time you say please as a child, you’ll probably start saying please more often. But suppose once in a while the candy doesn’t come. You might guess that you’d start saying please less often. After all, it’s not generating the reward as reliably as it used to.

But sometimes the opposite thing happens. It’s as if your brain, a born pattern finder, can’t resist the challenge. “There must be some additional trick to it,” murmurs your obsessive brain. You keep on pleasing, hoping that a deeper pattern will reveal itself, even though there’s nothing but bottomless randomness.”

– Jaron Lanier, “Ten Arguments for Deleting Your Social Media Accounts Right Now”

Just as algorithms are designed to organize and produce patterns, we are designed to spot them. Even when they aren’t necessarily there.

There is an inherent element of randomness to the recommender systems that run our feeds. These systems are constantly devising and running tests to determine what type of content provokes optimal engagement. In doing so, they occasionally throw a wrench in the machine, upsetting the established groove to see if a new direction is viable. We don’t always know when that’s happening.

The result is a whirlpool of half-blind guesses, flimsy cause-and-effect, and cyclical engagement cycles that trap users in a hamster wheel of confusion and perpetual motion. Though these A/B tests are conducted to refine the recommendation system, it is common for the randomization elements to collide with the organic spontaneity of the user’s own activity, resulting in a feed made up of content that may abstractly pique intrigue, but does not necessarily scratch any particular itch or deepen engagement with any particular topic.

This blind spot has a ripple effect that impacts consumers, creators, curators, and platform managers alike.

How Social Media Algorithms Weaponize Randomness

Think back to the last time you saw something on your social media feed that felt completely out of left field.

It’s odd enough to come across something that’s so jarringly separate from the type of content that’s normally curated for you, but what’s even odder is the comment section on videos like this — odds are, if you look to see what others are saying under posts that surprise you, they are equally perplexed as to why they’re being served this content. But they’re also engaging with it.

Chaos can catch like wildfire. These anomaly posts don’t just pop up and disappear like a blip on an individual’s feed; they manifest at random and catch compounding attention, more like an ever-expanding blackhole that sucks attention and engagement without providing any value to accompany their novelty. And that’s kind of the point.

In 2025, content creators deliberately start fights in their own comments sections. Influencers will intentionally spell a word wrong in a video just to get people to point it out. Entire accounts exist only to generate content that generates anger, conflict, or moral dissent. Because dissent requires engagement. And engagement is king.

One might assume, then, that a way to one up the algorithm is to refuse engagement entirely; to consume content quietly and passively, never letting it be known your real reaction to it. This is intuitive, but folly. Here’s the truth: The less you engage, the more important your engagement is.

That may sound empowering, but it really isn’t. Imagine you’ve been using Instagram for years, and have refrained from ever so much as liking or saving a post. You don’t share, you don’t comment. You don’t even follow. You log on, look at whatever is peddled automatically, then log off. Then, one day, you happen to drop a like on a photo of Snoopy. I hope it was worth it, because now, Snoopy is 100% of what the algorithm knows about you.

And Snoopy’s coming for you. Big time.

Press enter or click to view image in full size

Photo by Cris S. on Unsplash

Contrast this with the folks that are chronically online. Yes, people who engage with radical ideological consistency may find themselves in the narrowest of echo chambers, but as long as a person isn’t the exact stock image of a political or social archetype, consistent and varied engagement actually widens their social media landscape, prompting a wider diet of content than would exist if the algorithm had less to go off of.

Herein lies another reason that my experiment in randomness doesn’t work: I’m not genuinely engaged. I won’t go as far as to say that the algorithm is nuanced enough to understand when someone is feigning interest vs when they’re really invested in a topic, but it certainly knows not to weight my brief, passive consumption of impulse searches as heavily as the type of content that I consistently like, save, and share.

So is the real solution accelerationism? Is the way to beat the algorithm to overwhelm it? To comment on random videos and share everything you come across? To give it so much engagement — so much data — that it may as well have nothing?

Not really.

You’re Not Getting Away WIth It

While making an effort to actively engage with various forms of content can tick the needle in your quest for a more varied slate of feed decorations, it will not “defeat” the algorithm in the sense of nullifying its influence. Strategic, varied engagement may broaden recommendations slightly, but the platform’s models are resilient; the system will always extract signals from behavior.

No algorithm exists unto itself. No algorithm has desires, self-awareness, or anything else that makes our daily engagement with algorithms perilous. Algorithms exist only as tools for gathering, ranking, and categorizing data. Because algorithms themselves have no inherent bend towards one outcome or another, they are inevitably going to do more than those emplying them need or want them to do.

Think of a tech CEO as a farmer, and an algorithm as a magical creature the farmer has come across. This creature is incredibly efficient at harvesting crops, but only because it loves digging up and hoarding things. The farmer uses the creature to harvest crops, which is now done in record time, but the farmer now has to sort through an abundance of sticks, leaves, and buried bones/knicknacks that are dumped along with the crops.

Social media managers may want to know what videos you like. They may also want to know what videos make you angry, or scared, or confused enough to engage with long-term. But there’s no algorithm for that. All an algorithm can do is collect patterns. How long did you watch this type of thing? Is this type of thing meant to produce anxiety (horror content, inflammatory political content (i.e. “ragebait”), etc.)? Algorithms can accumulate data regarding your engagement with individual content, aggregate that data alongside the engagement of others, and then make decisions on what to show you and others. It is not responsible for categorizing you, others, or the content. That responsibility lies with its master. But here’s the catch: the master doesn’t have enough information to properly categorize anything.

It’s a paradox; the algorithm is the only one with enough of an understanding to properly pair consumers with content, but it is incapable of executive decision making of that sort.

Put more simply: Algorithms can detect patterns humans cannot, yet they cannot assign meaning or make judgment calls. The result is that the algorithm is uniquely capable of mapping engagement but incapable of responsible executive decision-making, while humans can make those executive decisions but cannot fully read the map the way an algorithm can.

The point is: You can’t outsmart the algorithm because it’s not actually as smart as you think it is. You can’t beat it at it’s own game, because it’s barely even playing.

The algorithm just is.

So… What now?

Press enter or click to view image in full size

Photo by Julius Drost on Unsplash

You Already Know The Solution

Sorry. You don’t want to hear it. I don’t want to say it. But until there is a fundamental restructuring of how tech companies collect and use data, there’s only one way out of this mess:

Log off.

In his book “Ten Arguments for Deleting Your Social Media Accounts Right Now”, author and Silicon Valley pioneer Jaron Lanier’s describes in beautiful detail how the inherent issue is not with social media itself, and that the internet, complete with all of the connection we feel we have gained from its most prominent sites, can absolutely be a force for good. Just not with its current systems of incentives and rewards.

At this moment in time, there is no meaningful way to game your algorithm. This isn’t to say that you’re forever doomed to live in an echo chamber, or even that you can’t engage with your social feeds meaningfully and thoughtfully — it is only to say that you are not and cannot be in control of how your feed of content is tinted.

The good news is that this is only inherent to algorithmic systems. Manually subscribing to and patronizing a variety of news outlets, creators, artists, and organizations directly, both online and in person, without the middle management of an algorithmic system (google included) remains a great way to ensure you have a wider context for your online activity.

This may have felt like a long winded way of saying “touch grass”, but maybe long winded can be good from time to time. Maybe it’s not a bad idea to spend longer periods thinking in directions we normally wouldn’t — in directions that aren’t being passively served to us.

Maybe we should stop trying to “beat” the things that have us so confused and start playing a different game entirely.