Fallingwater
The vibes can't be counted.
At some point in my professional life, everything came to be about data. Sometimes I hear a voice in the back of my head saying, “I’m talking about the data.” and I have no idea where it came from, or who it belongs to. Good data, clean data, actionable, governable data is everything in tech. I do know that. But in this modern moment exacerbated by generative AI and its many, many manifestations, we now have access to a near-constant data stream of our very lives. And not only that, but it’s starting to be repackaged and retold to us, forming narratives of our own experiences in ways we perhaps haven’t before evaluated.
From Spotify, I recently learned that I’ve listened to Maggie Rogers’ “Fallingwater” 222 times since its release in 2018.
Father John Misty’s “Nancy From Now On”? 224 plays, starting the day after my twenty-fifth birthday, apparently. Future Island’s “The Fountain,” 135 plays over the past six years. “Heat Waves” by Glass Animals, over two dozen times in a single weekend in 2022 (Lake Tahoe, a ski weekend frenzy, spring yawning into early summer, we played this up and down Highway 80 at ear-bleeding decibels).
If you haven’t explored this new generative playlist feature, let me enlighten you. You can now prompt Spotify via these preconstructed playlist templates, to craft a custom list of songs on a particular theme: “Roast My Listening,” “This Day in Music History,” “My Freeway Flight & Late Drives.” To develop these playlists, Spotify uses generative AI and personalization models to surface old listening patterns and contextualize them. This particular one is called “My Rediscover Weekly,” and it pulls songs you’ve listened to heavily but haven’t played in the past six months or so.
It’s an aural walk down memory lane, with each song accompanied by a blurb summarizing the data. Think: You were obsessed with this song on this one day in 2019, remember? For the span of three years, you played the motherloving shit out of this track. Is something wrong with you? Why have you listened to “Cold War” by Cautious Clay 553 times?
Just kidding. Spotify doesn’t judge like that (unless you prompt it to). The point of this playlist is to deliver the straightforward facts and figures of your musical meanderings, no more no less.
It was unsurprising to me that, sifting through this list, I could remember precisely what was happening in my life at the point each song spiked. There was a Perfume Genius track I submerged myself in starting in 2019, and it carried me all the way through a series of pandemic road trips in 2020. I remember entering Jackson Hole Valley after thirteen hours in the car with my cousin, blasting it out the window and feeling, for the first time in six months, that I could actually see a way out of this thing. A Zhu song I overplayed while home from college one summer. The truck I drove at the time had a real ace of a sound system, and I frequently blew out my eardrums in the small cab of it. On long enough drives and with good enough music, it started to feel like a sensory deprivation tank of my own making. I’d arrive at the pool I worked at or a friend’s summer barbecue in a near-perfect state of zen, my entire head slightly ringing.
Spotify doesn’t know that. The LLMs pulling this dataset don’t know, of course, the acute joy, fear, ecstasy, heartbreak, grief, whatever, that triggered the obsessive looping of these songs. Spotify wasn’t spiritually there the first time “Fallingwater” hit the AUX cord on a drive home from Coachella as my friend’s little sister threw up on the freeway shoulder but I didn’t really care, because I was mentally back at the Gobi Tent, watching that red scarf paint itself across the stage. It doesn’t know that performance broke me open to Humpty Dumpty proportions. (In fact, I would never be the same. So much so that I went home, bought the same pair of snakeskin pants Maggie was wearing (lol), and applied to grad school. It created, and I say this with no hyperbole, a BC/AD kind of demarcation.)
Since generating this playlist, I keep wondering, what do the numbers do to these memories? Do they complicate them, corroborate what I already knew or felt or guessed at? Are they changing the way I listen? Is there a risk of a musical Hawthorne Effect, where the mere act of being observed alters my behaviors, perhaps even my tastes? I don’t really think so. Was I at risk of forgetting these memories or songs had Spotify not resurfaced them for me? Probably not.
Data tracking doesn’t inherently “make” meaning. Repetition doesn’t necessarily signify importance—not really. If platforms like Spotify are trying to generate more traction by becoming the narrator of our personal histories through data and generative AI wrappers, exactly what is it they’re counting on us to care about? Are they counting on us to sift, as I have, through our own musical biography—one reliant on repetition alone? And how do I feel about a platform repackaging experiences and information in ways meant to elicit an emotional response like that? At best, is it a novelty? At worst, is it…kind of manipulative?
What they’re banking on is that high numbers are themselves a worthwhile story. And sure, I suppose they are. But they aren’t the only worthwhile story. And while other examples of these custom playlists rely on different prompts to create different outcomes—outcomes perhaps not as reliant on nostalgia but some other trendy gimmick—they all promise a legible version of ourselves, promise to help us name and quantify ourselves in a way we seem to be hungry for. It’s the phenomenon of Spotify Wrapped each year, everyone saying “wow this is so me.” Well of course, who else did we suppose listened to all those songs?
That data, however, can’t capture (let alone understand) the ineffable nature of a song itself, nor does it know why you gravitated toward it, what it might’ve carried you through, what part of your life that it scored. Perhaps one day Spotify will sync with your Gcal, your camera roll, (it can already sync with your texts), your fitness apps—and then, to be sure, it will try even harder to assemble a full picture of your life. I for one, don’t want to be around to see that, though something tells me I probably will be.
The Rediscover Weekly playlist is, despite much of what I’ve said here, harmless. If anything, it’s momentarily thrilling because it works like a mirror to a past self. And in a society suffering a crisis of meaning, perhaps we’re looking for ourselves everywhere. In 2012, we had an unlimited supply of Buzzfeed quizzes for that purpose; now, we have generative, customizable AI. Potato potahto. Know thy listening history, know thyself.
But honestly? 222 times for “Fallingwater”?
That number actually seems a little low.
Thoroughly enjoyed this! I’ve been thinking about the same topic a lot lately — and have Say It as my top listened to track of all time. My coming of age anthem boiled down to repeat plays, the rest of the tracks also marking distinct moments in my life with nuance I hope an algorithm can never really understand. Yes, harmless. But also, wish they’d invest in artists, music, life, etc instead of telling me what I already am pretty sure I know?