When was the last time you actually chose what to watch, read, or listen to?
Not scrolled until something caught your eye. Not clicked on what appeared in your feed. Not selected from a personalized row of recommendations. Actually chose—with intention, curiosity, or even just random whim—to seek out a specific piece of media.
If you’re struggling to remember, you’re not alone. Somewhere in the past decade, we stopped choosing our media and started accepting it. Recommendation algorithms haven’t expanded our options; they’ve quietly redefined what “choice” means, transforming it from an active process of discovery into passive consumption of whatever appears in our feeds.
The Illusion of Infinite Choice
The promise was intoxicating: every song ever recorded, every film ever made, every article ever published, all at your fingertips. Netflix replaced Blockbuster’s few thousand titles with an essentially unlimited catalog. Spotify offered 100 million tracks compared to the few hundred CDs you could afford to own. The internet gave you access to more writing than you could read in a thousand lifetimes.
But here’s what actually happened: most of us now consume a narrower range of content than we did before.
The algorithm’s job isn’t to show you everything available. It’s to predict what will keep you engaged, what will make you click, what will prevent you from leaving the platform. So it feeds you variations on themes it knows you’ll accept. You liked one true crime documentary, so now your Netflix homepage is a grid of murders. You listened to one lo-fi beats playlist while working, and Spotify has decided that’s your entire musical identity.
The paradox is that unlimited options have made our actual consumption more homogenous. When you had a shelf of 200 CDs, you chose from all 200. When you have 100 million songs, you listen to what’s served to you from the top of an algorithmically ranked list.
From Exploration to Optimization
There’s a fundamental difference between browsing and being fed. Browsing is inefficient, meandering, and open-ended. You might walk into a bookstore looking for a mystery novel and leave with a cookbook and a poetry collection because something on the shelf caught your eye. You might flip through radio stations and stop on a song you’ve never heard by an artist you don’t know, simply because something about it made you pause.
Algorithms have eliminated that inefficiency. They’ve optimized discovery right out of existence.
Every major media platform now operates on the same principle: don’t make users search, don’t make them choose, just show them what they’re statistically likely to engage with. YouTube’s homepage isn’t organized by category or topic—it’s a personalized feed based on your watch history. TikTok doesn’t even pretend to offer choice; you open the app and videos just start playing, one after another, each selected by an algorithm trained on your previous engagement.
This isn’t discovery. It’s prediction. And prediction, by definition, can only show you more of what you’ve already seen.
The Feedback Loop Prison
Here’s the insidious part: every interaction with an algorithm trains it to give you more of the same. Click on one article about a celebrity divorce out of idle curiosity, and the algorithm interprets that as deep interest in celebrity gossip. Listen to a song ironically, and it becomes part of your musical profile. Watch a video to see what the fuss is about, and the algorithm decides you want to see fifty more just like it.
Your choices aren’t just constrained by what the algorithm shows you—they’re constrained by your own past behavior, much of which was thoughtless, accidental, or context-specific. That video you watched at 2 AM when you couldn’t sleep? The algorithm doesn’t know you were sleep-deprived and clicking randomly. It just knows you engaged, and now that content defines you.
The feedback loop becomes inescapable. The algorithm shows you content based on your history, you consume that content because it’s what’s available, which reinforces the algorithm’s model of who you are, which determines what you see next. Your “choices” become a closed circuit, each selection narrowing the range of future options.
Breaking out requires active resistance—deliberately clicking “not interested,” aggressively pruning your watch history, or abandoning the platform entirely to seek content elsewhere. But that’s exhausting, and most people don’t bother. It’s easier to accept what you’re given.
The Manufacturing of Taste
We like to think our preferences are intrinsic, that we gravitate toward certain types of music or films because of who we fundamentally are. But taste is significantly shaped by exposure, and algorithms control exposure with unprecedented precision.
Spotify doesn’t discover what music you like—it determines what music you’ll hear often enough to develop a preference for. If an artist appears in your Discover Weekly playlist three times, in a Release Radar once, and gets slotted into several algorithmically generated playlists you listen to, you’ll start to feel familiar with their sound. That familiarity breeds preference. Soon you’ll think you “like” that artist, when really, the algorithm simply ensured you heard them more than the thousands of artists you never encountered.
The same dynamic plays out across platforms. Netflix’s algorithm doesn’t just recommend shows you might enjoy—it shapes what gets made in the first place. Creators and studios now optimize content for algorithmic success, creating shows with hooks in the first episode, cliffhangers at predictable intervals, and familiar genre markers that fit neatly into recommendation categories. Your taste for this content isn’t natural; it’s cultivated by an ecosystem designed to produce algorithmically legible media.
What We’ve Lost
The shift from choosing to accepting has costs that go beyond individual media consumption. We’ve lost the ability to be surprised by our own interests, to discover that we enjoy something we would never have predicted. We’ve lost the productive friction of encountering art or ideas that don’t fit our established preferences.
We’ve also lost a certain kind of cultural literacy. When everyone’s feed is personalized, we stop having shared cultural experiences. You and your friend might both “watch Netflix,” but you’re essentially watching different services with minimal overlap. There’s no modern equivalent of everyone talking about the same TV show because everyone saw it on Thursday night—we’re all watching different things at different times, served to us by algorithms trained on our individual behaviors.
Perhaps most importantly, we’ve lost agency. The ability to choose—really choose, with intention and purpose—what media to consume is a form of self-determination. When that choice is outsourced to an algorithm, we become passive recipients rather than active participants in our own cultural lives.
Reclaiming Choice
The solution isn’t to abandon these platforms entirely—that’s impractical for most people, and the algorithms do occasionally surface genuinely good content. But we need to recognize what’s happened and actively resist it.
That means deliberately seeking out media through non-algorithmic means: reading recommendations from actual people, browsing physical or digital spaces without personalization, following your curiosity even when it leads somewhere unexpected. It means treating algorithmic recommendations with skepticism rather than trust, recognizing them as predictions based on limited data rather than accurate reflections of your interests.
Most importantly, it means reclaiming the right to not know what you want. Algorithms can’t handle open-ended exploration; they always need to categorize and predict. But some of the best cultural experiences come from wandering without a destination, from being open to surprise, from choosing something for no better reason than that it seems interesting in the moment.
Choice isn’t just about selecting from available options. It’s about the freedom to explore, to change your mind, to discover things that don’t fit neatly into your established preferences. Recommendation algorithms haven’t given us more choice—they’ve replaced choice with prediction, exploration with optimization, and agency with acceptance.
The question is whether we’re willing to take it back.
