Breaking Away From the Algorithm Is a Myth. That’s Not the Bad News It Sounds Like.
From soap opera cliffhangers to TikTok’s infinite scroll — the hook was never new, only the exit has disappeared.
I’m old enough to remember a version of life before any of this existed. You wanted to know what was good, you asked a person, in a room, and you watched their face while they told you. Nobody was optimising that conversation for anything. Five decades in, watching a generation grow up talking to a feed instead of a face, it’s fair to ask the blunt question: have we actually evolved, or have we just gotten quieter about not thinking for ourselves? That question is worth sitting with before reaching for the easy answer, because the easy answer — “we used to be free, now we’re not” — isn’t quite right either.
Here’s the fantasy: somewhere under all the recommendation engines, there’s a “real you” waiting to be dug out. Delete the apps, kill autoplay, and your true taste surfaces again, like it was buried, not erased. I don’t buy it. There was no pre-algorithmic self sitting around waiting for rescue. Taste has always been manufactured by something outside you — critics, radio DJs, whoever got to decide what went at eye level in the record shop, whoever in your friend group had the social capital to make something cool. The algorithm didn’t invent influence. It just got faster at it, and better at hiding what it’s doing.
Slave to the algorithm
Forget content for a second. The real complaint is architectural. These systems are built to maximise engagement, full stop, and engagement is not the same thing as your wellbeing. A friend who tells you about a song wants you to like it. A feed just wants you to keep opening it. So when people say they feel enslaved to the algorithm, that’s not really a metaphor. It’s a fairly literal description of an incentive mismatch: something is being optimised, and it isn’t you.
Breaking away
But “escape” has a hole in it. There’s no neutral, untouched ground you land on after quitting Instagram. You don’t rediscover some pure inner voice; you just fall back on whatever else is available: family, print, your own patchy memory of what you used to like before all this. Those are still algorithms. They’re just slower, and run by humans instead of servers. Somebody proud of having “broken away” because they now only read physical books hasn’t actually left curation behind. They’ve swapped one set of curators for another. That’s a real upgrade. It’s just not an escape.
Where this pattern actually came from
None of this started with an app. It’s worth tracing the lineage, because it shows the hook isn’t new — only the exit has been removed from it, one generation at a time.
Television soap operas built entire episodes around the cliffhanger: end mid-tension, make the audience wait a day or a week, and let the wait itself do the work of bringing them back. The format wasn’t a one-off trick — it got reproduced independently across British soaps, American daytime drama, and telenovelas around the world, because it exploits something basic about how curiosity works, not something specific to any one culture.
Advertising formalised a related idea. AIDA — Attention, Interest, Desire, Action — was formulated by advertising pioneer Elias St. Elmo Lewis in 1898, built around a funnel with an exit: get someone’s attention, build interest and desire, then get them to act — buy the product, and the loop resolves.
Clickbait strips the funnel down to a headline. It works by opening what psychologists call an information gap — a felt discrepancy between what you know and what you want to know. George Loewenstein’s 1994 paper on the psychology of curiosity described curiosity itself as a kind of deprivation, a state that builds precisely because a knowledge gap has become impossible to ignore, pushing a person to chase down whatever would close it. “You won’t believe what happens next” isn’t information — it’s an engineered gap. And because there’s no purchase to complete, no funnel to exit, the gap just reopens with the next headline.
Infinite scroll removes the exit at the level of the interface itself, not just the copy. The feature was introduced by designer Aza Raskin in 2006, originally meant to smooth out page-loading, before becoming the default engagement mechanic of nearly every major platform. It works on a principle borrowed straight from slot machines: rewards that show up unpredictably rather than every time train the strongest, hardest-to-quit habits of all. You don’t scroll because you’re always rewarded. You scroll because you might be.
The economics changed too, not just the psychology
Television and radio paid enormous sums to produce the content that built an audience, then sold that audience to advertisers. Platforms inverted the model. YouTube’s standard revenue split gives creators 55% of ad revenue on long-form video and 45% on Shorts, with the platform keeping the rest — but the creators are the ones supplying the labour, equipment, and time that television networks used to fund directly. TikTok’s effective creator payout is widely reported to run closer to 10–20% of ad revenue, a much thinner cut for the same unpaid production labour. Either way, the “ratings” that once came from Nielsen boxes have been replaced by likes, views, and watch time, generated for free by the public, and sold to advertisers at a fraction of what broadcast-era content cost to produce.
If the audience has effectively become unpaid production labour, the next question is whether that arrangement holds, breaks, or gets regulated — and that’s already in motion. The EU’s Digital Services Act now requires large platforms to explain how their recommendation systems rank content and to offer a non-profiling feed option where feasible, and in 2026 a Dutch court ordered Meta to provide a more durable opt-out to a purely chronological timeline rather than its algorithmic default. Neither undoes the underlying economics, but both are concrete signs that the “demand transparency” counter-movement is already underway, not hypothetical.
So is it possible to walk away?
Structurally, yes: people delete the apps and stay off them. But whether it’s a fair fight is a different question, and the honest answer is no. Infinite scroll, autoplay, and variable-timed rewards aren’t accidents of good UX — they’re deliberately engineered and continuously tested by teams whose job is to raise the number. Willpower against a product built by people paid to defeat it isn’t a level contest. That’s a sharper version of the “no neutral ground” point above: it’s not just that there’s no escape to a pure state, it’s that the loop you’re trying to leave was built by adversaries to your attention in a way a bookstore shelf, a radio DJ, or even a soap opera’s writers’ room never quite was.
Where this might be heading
Fiction got here before the trend lines did. One well-known dystopian TV episode imagined a world where a single aggregated score, built from peer ratings after every interaction, decided your access to jobs, housing, and social standing. It’s fiction, but it’s not pure fantasy — pieces of it already exist, just scattered rather than unified: a credit score here, a driver rating there, a follower count somewhere else, each with real but partial consequences.
The more plausible ten-year trajectory isn’t one number ruling everyone’s life — platforms have commercial reasons to keep their scoring systems proprietary and separate, not merge them into a single dystopian dial. What’s more likely is quieter and messier: today’s engagement scores mostly cost you visibility, but the real question is whether they start costing you access — insurance terms, hiring decisions, loan rates — the way credit scores already do for a narrower slice of behaviour. It’s also likely that a growing share of what fills these feeds stops being human-made at all, generated instead to hit the same engagement targets, which would break the “audience-as-unpaid-labour” model this piece has been describing, since the platform wouldn’t need anyone to produce anything — just to keep watching. And the transparency push described above — regulation, court rulings, chronological opt-outs — will likely keep expanding rather than fading, even as it runs up against platforms’ incentive to route around it.
None of that is a unified prophecy. It’s closer to: expect more of what’s already happening, in more places, with higher stakes attached, and a parallel market growing around the discomfort of it.
Where that leaves you
You were never going to be some sovereign, uninfluenced chooser. Nobody was, ever. So the fight isn’t for purity — it’s for which loop you’re in, and how much say you get inside it, and whether the people running that loop are being paid to help you or to keep you there. “Breaking away from the algorithm,” properly understood, doesn’t mean freedom from influence. It means trading a fast, opaque, engagement-hungry loop for a slower one you can actually see the workings of and steer. That’s a real difference. It’s just a smaller, less romantic one than the story we usually tell about it — and it’s the difference that will decide which version of the next ten years we end up living in.
Sources
Loewenstein, G. (1994). The Psychology of Curiosity: A Review and Reinterpretation. Psychological Bulletin. Summary: https://psychologyfanatic.com/information-gap-theory/
Collins, G. “Why the infinite scroll is so addictive.” UX Collective: https://uxdesign.cc/why-the-infinite-scroll-is-so-addictive-9928367019c5
“YouTube Partner Program, Explained.” YouTube Blog: https://blog.youtube/creator-and-artist-stories/youtube-partner-program-explained/
“YouTube Revenue Sharing: How the 55/45 Split Really Works” (includes TikTok comparison): https://ytmoneycalculator.com/blog/youtube-revenue-sharing/
“E. St. Elmo Lewis and the origin of the AIDA model” (1898): https://www.toolshero.com/toolsheroes/elias-st-elmo-lewis/
“36 Predictions for Social Media Marketing in 2026” — on the Dutch court ruling requiring Meta to offer a chronological feed opt-out: https://www.socialmediatoday.com/news/36-predictions-social-media-marketing-2026/802195/
“Algorithmic Transparency Laws 2026: What Platforms Must Disclose” — on EU Digital Services Act recommender-system disclosure requirements: https://beyondtmrw.org/article/algorithmic-transparency-laws-what-platforms-must-disclose-in-2026