Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist by Liz Pelly. Atria/One Signal Publishers, 288 pp., $29.
The release party for Mood Machine took place in the beginning of January 2025, at the theater at St. Joseph’s University in Brooklyn. The author, Liz Pelly, whose book argues that the streaming music giant Spotify takes advantage of artists, surveils subscribers, funds AI warfare, and forced upon us the popularity of “chill” muzak, was interviewed by Max Alper, a radical composer who used to post under the handle @la_meme_young. It was copresented by The Baffler, where much of Pelly’s reporting into Spotify’s exploitative practices was first published, and United Musicians and Allied Workers, an upstart union that is leading a campaign to pass legislation that would increase streaming royalty rates, was on hand to table.
It’s hard to imagine that Spotify’s top brass would have agreed with much of what was said either in the book or at the event. Which is why it was so strange when, less than two weeks later, the company threw a party that, in a way, seemed designed to confirm all of it, hosting right-wing podcasters Ben Shapiro and Tim Pool, as well as MAGA country singer Riley Green, at brunch in DC a day before Donald J. Trump’s second inauguration. I doubt that Pelly was invited, but if the goal was to convince skeptical readers that Spotify is indeed run by unscrupulous opportunists unbothered by the trail of destruction left in their wake, then its organizers should consider their soiree a complete success. They even donated $150,000 to Trump’s inauguration committee for good measure.
First Spotify uses data to get subscribers to listen to terrible music, then it takes subscribers’ listening data and sells it to terrible people.
Mood Machine, the year’s most celebrated work of long-form music criticism, asks two questions, one historical and one critical: How did we get here? And what are the consequences? The story begins in Sweden. Spotify was founded there in 2006, but Pelly opens five years earlier, when 15,000 antiglobalization protesters took over the streets of Gothenburg outside the 2001 European Union summit. The demonstrations were violently suppressed, and in their aftermath, many of those involved took up the cause of copyright abolition, envisioning the internet as a twenty-first-century commons where ideas and information are transferred freely and private property rights don’t apply. If you’ve ever illegally downloaded a movie, these are the people you ought to thank. File sharing, to them, was a political act, and they formed a research group called Piratbyrån to theorize and expand the practice. One of their projects was the torrent tracker known in English as the Pirate Bay.
Why did Sweden, of all places, become such a hub of anticapitalist bootlegging? According to Pelly, the answer is Nordic social democracy. In Sweden, the internet is treated as a public utility and music lessons are subsidized by the government. It would be hard to grow up in such an environment and not resent the Lars Ulrich–fronted overreaches of the Amerikansk entertainment business. By the mid-2000s, torrenting had become so popular there that the recording industry considered the entire country to be a “lost market”—a turn of phrase that David Graeber would have loved.
Piratbyrån’s Peter Sunde remembers Spotify founders Daniel Ek and Martin Lorentzen as “advertising dudes who wanted to sell more advertising and realized that music was cheap.” Early on, they pitched themselves as Pirate Bay fellow travelers: They purchased uTorrent (a similar file-sharing client) and hired comrades from Stockholm’s anarcho-coder community. They even populated Spotify with torrented music, which explains why—as millennials reading this may remember—its original database was filled with needle drops and labeling errors. When Ek and Lorentzen dealt with the music industry, on the other hand, they framed their mission in Thermidorean terms, billing their product as the way to get people paying for music again.
This shamelessness set them up for success in the world of tech, but they were missing one thing that every start-up needs: a nepo baby. Help arrived in the form of Fred Davis—son of RCA Music Group CEO Clive—a lawyer who had previously consulted for Napster and Kazaa on the way to becoming, in the words of Billboard, “music’s leading financial gatekeeper,” an enclosure metaphor that is all too apt. Meanwhile, the Motion Picture Association of America lobbied the US government to impose sanctions on Sweden unless the country shut down Pirate Bay. Sweden complied, sending Sunde to jail and clearing the runway for Spotify’s takeoff. By the fall of 2008, the three major record labels—Sony, Universal, and Warner, which had gone so far as to hire the cop who led the case against Sunde—had bought in to Spotify’s vision, leveraging their back catalogues to lock down ownership stakes. Meanwhile, the radical tech workers who helped launch the company were left disillusioned, like so many radical tech workers before and since.
If music is a utility, then one path forward would be to socialize its production and consumption, de-enclosing it.
As the narrative approaches the present day, Mood Machine’s chronological progression slows. The rest of the book’s chapters are organized around overlapping issues—appropriation, payola, low compensation—that affect the artistic and financial lives of working-class musicians. For consumers, the story is almost as bleak: First Spotify uses data to get subscribers to listen to terrible music, then it takes subscribers’ listening data and sells it to terrible people. The book covers banking start-ups that have used potential borrowers’ streaming stats to make decisions on loans, as well as Ek’s €100 million investment in a German military contractor that specializes in “battlefield AI.” (He threw in an additional €600 million in June.)
“wage war on nostalgia and bourgeois taste”
Coming from a punk background, Pelly is particularly concerned with what she calls the “conquest of chill.” A Spotify employee, one of her sources, recalls Ek telling staff that the company’s only real competitor is silence. Trying to beat silence at its own game, it promoted a range of passive-listening playlists that displaced quiet without making any demand on the listener’s ears. This music has a numbing effect, by design. “The conquest of chill reflects an industry content to profit from a world of disconnection,” Pelly writes. “Capitalism both alienates us and sells us tools to distract us from the loneliness of nonstop alienated labor.” Yes, that does seem to be the case—though I’m sometimes willing to look past it if capitalism’s distraction music goes hard enough. In this instance, however, it stinks. Spotify refined its specific chill aesthetic by hollowing out an existing, organically chill genre called “lo-fi beats.” Using the financial incentives that come with being featured on its in-house playlists the company rewarded the lo-fi beats producers who outblanded their peers, turning a creative community into a repository for milquetoast study music. Now users are encouraged to invert the original rock ’n’ roll dream and lo-fi around the clock, letting the “Wake Up Lo-Fi Mix” lead to a “mellow lofi morning” before tuning in to the “Walking Lo-Fi Mix” for a stroll to the “lofi cafe,” or perhaps a bit of gardening in the “lofi garden” under “lofi summer haze.” Later, the “Cooking Lo-Fi Chill Mix” is followed naturally by the “Dinner Lo-Fi Mix,” “Cleaning Lo-Fi Mix,” and some “Late Night Beats.” Then it’s off to “lofi sleep” and the lo-fi day begins anew. (Every phrase in quotation marks is a real Spotify playlist.)
A similar arc befell the genre of hyperpop, which Pelly describes as “a shorthand for a post-SOPHIE, post–PC Music internet aesthetic, a more queer and trans alternative to mainstream experimental electronic music culture, often with pitch-shifted vocals, drawing influence from vaporwave, nightcore, chiptune, J-pop, Y2K pop, hip-hop, and all sorts of dance music.” If that’s not tracking for you, find someone with really baggy pants and ask them nicely to explain. It’s basically glitchy music that’s impossible to study to—which is exactly the point. The sound was an explicit rejection of Spotify’s forced tranquility. The hyperpop producer umru tells Pelly that his artistic vision is indebted to a tweet by another hyperpop producer named Iglooghost. The tweet said, “NO TIME FOR ‘CHILL VIBES.’ … IT IS 2K16 LIKE CMON SPEED IT UP MY GUY !!!!” umru printed it out and hung it above his desk.
Spotify originally showcased this music in a playlist called “Neon Party.” In 2019, the editors renamed the playlist “Hyperpop” after seeing the term appear in an ever-evolving map of microgenres maintained by the company’s resident “data alchemist.” As with the lo-fi beats playlists, placement on “Hyperpop” came with a financial windfall, leading to a large bump in streaming revenue and contracts with major labels. Nearly overnight, Spotify became one of the main arbiters of success in a scene that was to a large extent created in rebellion against it. This had predictable consequences. According to a 2022 interview with the recording artist quinn—who two years earlier had been the star of a New York Times puff piece about the playlist’s impact—Spotify’s intrusion led to the elevation of straight, cis, white men over the predominantly queer creators of the genre. It got bad enough that one member of the scene made a thirty-minute YouTube documentary to prove the thesis that, as stated in the video’s title, “Hyperpop was not invented by Spotify in 2019.” “Why does it feel like we were all erased?” asks one artist who’s featured.
When Ek and Lorentzen dealt with the music industry, on the other hand, they framed their mission in Thermidorean terms, billing their product as the way to get people paying for music again.
Spotify was born by co-opting the ideas and energy of Swedish radicals, and it quickly became an instrument for intensifying the co-option of radical currents within the music industry. For artists like those making hyperpop, the platform offers a double-edged sword. To be included is to have your work incorporated into something you oppose, but to be left out is to lose access to one of the few income streams musicians have access to. It’s a new version of a dilemma Michael Denning described in his 2010 essay “Wageless Life”: “Under capitalism, the only thing worse than being exploited is not being exploited.”
No one in Mood Machine understands this better than the jazz musician who signed a one-year contract to record instrumentals for a third-party start-up that uses fictitious artist profiles to game the algorithm and further obscure the financial arrangement that underlies its tracks. Reading his anonymous quotes, I imagined him cast in shadows, speaking in the pitch-shifted tone that used to be applied to victims of horrible crimes in episodes of Unsolved Mysteries. “Nobody I know would ever go into the studio and record music this way,” he says. “And yet, hundreds of songs are being made and going on these playlists all the time that are exactly like this.” The biggest player in this cottage industry is Epidemic Music, a Swedish company backed by some of the same venture capital firms that elevated Spotify. These companies take complete ownership of their recordings, going so far as to force their composers to resign from the organizations through which songwriters collect royalties. In return for signing sweetheart deals that reduce their royalty rate, they are awarded prime placement on Spotify’s playlists. For the anonymous jazz musician, this proletarianization has been “pretty much joyless.”
Many of the good Marxist terms apply here. “Alienated labor”—that’s obvious. “Reserve army of labor”—look around; the music industry has a big one. “Real subsumption”—meaning that the very process of making music is reorganized to maximize profit. And, again, “enclosure,” the privatization of a shared good and the severing of the community ties that had been attached to it. Pelly has a keen eye for this, and she’s troubled by what she sees as a threat to the practice of communal music-making (what some might call “jamming”) as well as the breakdown of a long-held bond between musician and listener. “The relentless promotion of anonymous producers,” she writes, “seems to be part of a larger effort that aims to disconnect listeners from the makers of the music they’re consuming, laying the groundwork for users to accept the hyper-normalization of music made using generative AI software.”
Maybe someday these tracks will become cult favorites among a new generation of music nerds who find this whole thing fascinating and quaint. Indeed, when Mood Machine starts getting into AI, the saga of the anonymous jazz musician already does seem quaint. Spotify is now home to countless hours of music that’s computer generated, but where AI has most taken hold is in the app’s recommendation algorithm. In the company’s early era, this was conceived as a tool for expanding listeners’ tastes by introducing them to artists and genres they had never heard before. However, the employees who advocated for this vision—for instance, aforementioned “data alchemist” Glenn MacDonald, who after being laid off in 2023 published a defense of streaming titled You Have Not Yet Heard Your Favourite Song (2024)—were marginalized and pushed out as personalization replaced exploration as the rule of the day.
A Spotify employee recalls Ek telling staff that the company’s only real competitor is silence. Trying to beat silence at its own game, it promoted a range of passive-listening playlists that displaced quiet without making any demand on the listener’s ears.
Spotify’s recommendation capabilities have been further diluted by a feature ironically called “Discovery Mode.” Through the company’s “Amplifying Artist Input in Your Personalized Recommendations” program, artists and labels can buy algorithmic priority in exchange for a 30 percent reduction in royalties. A feature called “Pre-Campaign Insights” even uses AI to process tracks and recommend the ones to promote, further narrowing the range of music that’s actually discovered on the platform. Discovery Mode, which made €61.4 million in its first year of implementation, puts further downward pressure on musicians’ earnings: As more and more submit to pay cuts in exchange for exposure, the algorithm becomes increasingly biased against the ones who refuse.
On the user side, these changes have accumulated into a textbook example of enshittification. Many Spotify users I talk to agree that at this point the platform has reduced the amount of new music they listen to rather than expand it. Faced with an increasingly convoluted interface and an AI autoplay engine driven by the feedback loops of the listener’s pre-established habits, it’s often easiest to press play on an old favorite and not venture further. Sometimes, navigating Spotify, I feel like I’m in a Truman Show of my own taste. A friend recently tried to convince me of the merits of a 1980s synthpop style called “Hi-NRG.” To this end, they sent me the official Spotify playlist “Hi-NRG Mix.” What they didn’t realize was that this playlist is “personalized” via AI: When I opened it on my account, it started playing country music, something that Spotify is obsessed with recommending me. As of 2023, Spotify employed over seven hundred people to get these personalized playlists just right. “Creepy?” I wrote in the margin next to this piece of information. I’m not sure how I feel about a team many times the size of the New York Philharmonic fine-tuning a formula of Lucinda Williams and Willie Nelson hits with which to drip-feed me throughout the day. Actually, I take that back. I hate it.
Spotify was born by co-opting the ideas and energy of Swedish radicals, and it quickly became an instrument for intensifying the co-option of radical currents within the music industry.
So what can we do to change things? Pelly’s conclusion, the longest chapter in the book, adds this question to Mood Machine’s inquiry. After a couple hundred pages of unremitting critique, it’s remarkable how optimistic she comes out at the end. It’s as if finding out everything that has gone wrong has only deepened her conviction that things can instead go right. United Musicians and Allied Workers, the organization that tabled at her real book launch, figures prominently in this vision. Its Living Wage for Musicians Act, which was introduced to Congress last year by Rashida Tlaib, would create an additional royalty stream paid directly from streaming platforms to musicians.
Legislation that protects workers and funds the arts? It wouldn’t be the worst thing that could happen to this country. Unfortunately, while many in America long for a state more like Sweden, Sweden seems to be becoming more like America. Just before Pelly’s reporting trip in the fall of 2023, the country’s right-wing government slashed funding for both adult education (an important income source for working musicians) and kulturskolor, the free cultural schools where children learn to play instruments.
On a more local level, Pelly, befitting her DIY roots, finds inspiration in a cooperatively owned streaming service called Catalytic Soundstream. At the end of each month, half of its revenue goes toward maintaining the site (and paying for extras like editorial content) and the other half is split evenly among the changing list of thirty artists whose work is showcased. Public libraries in places like Seattle and Ann Arbor have created their own streaming services to promote musicians working in their cities. Anyone with a library card can listen, and artists are compensated with funds procured by the institution. The scale of these ventures may be small—Spotify has many more lo-fi playlists than Catalytic Soundstream has artists—but they point the way toward how music might be distributed under a more just economic system.
In Mood Machine’s intro, Pelly criticizes Spotify for valuing music “more as a utility than as art.” Fair point. But maybe this observation contains the seed of some future dialectal inversion: If music is a utility, then one path forward would be to socialize its production and consumption, de-enclosing it. This would require a change that reaches far beyond the music industry—if the cost of housing continues to rise, for instance, then a few extra decimal points per stream won’t provide enough income for the average songwriter to pay their rent. But Mood Machine’s most important idea might be that the crisis of music—and art in general—isn’t something that should remain on the back burner until the larger issues are figured out. Music, at least the best of it, offers its own vision of what a life in common might actually be like. It has long been thought that early humans learned to speak to each other in part by trying to sing the songs of birds, and that they developed new forms of communal expression by playing melodies on flutes made from the animal’s hollow bones. In more recent times, music has heralded—and in some ways prefigured—all sorts of social change. Who’s to say what possibilities lie in the sounds being made today? By freeing this work from the claws of VC vultures, we may yet find our way to a world where we can hunt in the morning, fish in the afternoon, rear cattle in the evening, and make playlists after dinner, just as Marx had in mind.