The DJ Chose the Song and You Loved It Anyway. Now the Algorithm Knows Your Taste and You're Bored.

Somewhere in America in 1993, a kid was driving to school with the radio on. A song came on she'd never heard before. It wasn't quite what she would have chosen — a little too slow at first, a little unexpected — but the chorus hit and she reached over and turned it up. By the time she got to school she was humming it. By Friday she'd bought the cassette single at Kmart.

That's how discovery used to work. Someone else made a choice, and you encountered it whether you were ready or not. You didn't curate the moment. The moment arrived, uninvited, and sometimes it changed something.

Now open Spotify. You have access to over 100 million songs. The app has been studying your listening habits for years. It knows your tempo preferences, your genre patterns, your time-of-day tendencies. It has built you a personalized radio station more sophisticated than anything that existed in 1993.

And yet a lot of people will tell you, quietly, that music doesn't feel as exciting as it used to.

What the Radio Actually Was

American radio in its commercial peak — roughly the 1960s through the 1990s — was a shared cultural infrastructure. A Top 40 station in Cleveland played the same songs as one in Phoenix. When a song crossed over, it crossed over everywhere, more or less simultaneously. The charts reflected something real: this is what the country is listening to right now, together.

The DJ was a gatekeeper, and not always a benevolent one. Payola scandals, format rigidity, and the slow consolidation of radio ownership under companies like Clear Channel meant the system had serious problems. Not every great song got airtime. Plenty of mediocre songs got played to death. The gatekeepers had their own interests.

But gatekeepers also created shared experience. When "Smells Like Teen Spirit" started getting played in 1991, you didn't have to seek it out. It found you. It found everyone. That simultaneity was part of what made cultural moments feel like moments — because you and your classmates and your older sister and the guy at the gas station all heard it at roughly the same time and had something to talk about.

The water cooler conversation about last night's song required that everyone had access to the same water.

The Personalization Promise

Streaming arrived with an appealing pitch: no more being held hostage to someone else's taste. No more sitting through songs you hate. No more format restrictions. Just music, matched to you, on demand.

The pitch was honest. Streaming genuinely delivered what it promised. Spotify's Discover Weekly, released in 2015, was legitimately impressive — a weekly playlist generated by algorithm that introduced users to music they hadn't heard but actually liked. It felt like magic. It was, in a technical sense, magic.

But a funny thing started happening as the personalization got better. The more precisely the algorithm mapped your preferences, the smaller the circle of music it showed you. It wasn't showing you what you might love. It was showing you variations of what you already loved. The edges got sanded off. The surprises got optimized away.

There's a term in recommendation system design for this: the filter bubble. The algorithm is trying to minimize the chance you'll skip a song, because skips are bad for engagement metrics. The safest way to avoid a skip is to play something close to what you've already approved. Over time, your listening profile becomes a mirror, not a window.

The Serendipity We Didn't Know We Needed

There's a specific kind of discovery that radio and MTV enabled that algorithms are structurally bad at producing: the accidental encounter with something genuinely outside your existing taste.

In 1985, a kid who only cared about heavy metal might flip past a channel and catch a Prince video and have his entire sense of what music could be completely rearranged. That didn't happen because he sought it out. It happened because he was a captive audience to a shared broadcast, and something unexpected landed.

The opt-in nature of streaming makes that kind of encounter nearly impossible. You choose what to play. Even when you choose a "discovery" playlist, you're choosing it — which means you've already self-selected for openness to discovery, which is very different from being surprised by something you didn't ask for.

Musicians and critics have noticed the downstream effects. The album as a format has weakened significantly, because streaming encourages listening by individual track rather than as a cohesive experience. Attention spans for new music have compressed. Songs that don't hook within the first fifteen seconds get skipped, so production has shifted to front-load the emotional payoff. The medium shapes the art.

The Paradox of Everything Available

Psychologists who study decision-making have documented something called choice overload — the phenomenon where having more options leads to less satisfaction with whatever you choose. The classic jam study, conducted by researchers at Columbia and Stanford, found that shoppers presented with 24 varieties of jam were far less likely to buy any jam at all compared to shoppers who saw only 6 options.

Music streaming is a jam study with 100 million jars.

When you had a cassette with ten songs on it, you listened to all ten songs. You didn't have the option to skip to something better. You gave each song the time it needed to become familiar, and familiarity is most of what we call love in music. The songs you know best are usually the songs you were forced to hear enough times to internalize.

Now, with infinite alternatives one tap away, the patience required for that kind of listening is genuinely harder to sustain. Why sit with a song that doesn't immediately grab you when there are 100 million others waiting?

What We're Actually Missing

The honest answer isn't that music is worse. By most technical measures, production quality is extraordinary. There are more artists making more music in more genres than at any point in history. The sheer volume of genuinely good music available today would have seemed impossible in 1993.

What's different is the cultural context around the music. Shared listening — the experience of hearing the same song at the same time as everyone else — has fragmented into millions of individual playlists. There's no song right now that everyone in America is hearing. The charts exist, but they don't reflect a shared experience the way they once did. You can be deeply into music and have almost zero overlap with what your coworkers are listening to.

That fragmentation has a cost. Music has always been partly about belonging — to a moment, to a generation, to a place. When everyone's listening to something different, the music can still be great, but the belonging is harder to find.

The kid in the car in 1993 turned up a song she didn't choose, and it became part of her. She didn't need an algorithm to tell her she'd like it. She just needed someone to put it on.