Tech Won't Save Us - How Spotify Remade the Music Industry w/ Liz Pelly

Episode Date: February 6, 2025

Paris Marx is joined by Liz Pelly to discuss how Spotify changes how we listen to music and the broader impacts it has on the wider music industry.Liz Pelly is a music journalist and the author of Mood Machine.Tech Won’t Save Us offers a critical perspective on tech, its worldview, and wider society with the goal of inspiring people to demand better tech and a better world. Support the show on Patreon.The podcast is made in partnership with The Nation. Production is by Eric Wickham.Also mentioned in this episode:You can read an excerpt of Liz’s book in Harper’s.The CEO of Suno AI said people “don’t enjoy” making music.Support the show

It's very risky when we start letting companies like Universal Music Group and these big tech companies be determining who is a serious artist and who is an unprofessional hobbyist who doesn't deserve to make any royalties at all for their work. These are really slippery slopes, I think. Hello and welcome to Tech Won't Save Us, made in partnership with The Nation magazine. I'm your host, Paris Marks, and this week my guest is Liz Pelley. Liz is a music journalist and the author of Mood Machine, The Rise of Spotify, and The Cost of the Perfect Playlist. Now, Liz has been on the show in the past. When I was preparing for the show, I was imagining, you know, it was probably a couple of years ago, like maybe 2022. But then when I got on the call with Liz, she reminded me that actually she was on way back in 2020 in the first year of the show to talk about Spotify with me. And on one hand, I couldn't

believe it was that long. But on the other, it was a reminder of how long I've actually been making Tech Won't Save Us and how we're coming up on the fifth year anniversary of this show. So I was super happy to finally have Liz on after all of those years, apparently, to go through her new book and to discuss these really important things that she found in reporting out this story about Spotify, its impact on artists, and its impact on the way that we listen to music. The first episode we did this year was with Will Tavlin talking about Netflix and the impact of the Netflix model on film and television. So I think that this is a really good accompanying episode to a conversation like that, right? Instead of looking at film and TV, looking at the music industry and, you know, what this kind of streaming model and particularly how Spotify has implemented it

has really meant for music and, you know, how we consume and how we experience this really important art form. I'm sure you've heard about how Spotify, you know, doesn't pay artists particularly well. We obviously go over that in this conversation, but there's a much deeper story to be told here about how Spotify works, about how it impacts music, about, you know, the impacts of, you know, this model of collecting all this data on us and, you know, supposedly making music and playlists and all this kind of stuff that is related to data. But Liz talks about the much deeper consequences of that entire shift and how, you know, it is obviously not served by improving the music listening experience. Spotify was originally created not by people who were

concerned about the music industry, but by people who were trying to build an advertising business. And that has affected the way that they treat this platform ever since that beginning moment. I was particularly thrilled to talk, you know, briefly at the end about what some alternative models would look like, how we might approach this in a different way if things weren't just shaped by, you know, the need to maximize revenue and the power of particular companies over this art that we listen to and how we can imagine different ways of recontextualizing and experiencing music in a different way. So, you know, I don't want to spoil that, especially if you haven't read the book yet. But, you know, obviously, if you listen

to our interview, you'll find out about it in just a little while. You know, at this point, streaming does feel very normalized, right? It's been around for long enough that many of us have streaming services, whether they're for video or for audio. And it just feels that this is how we need to consume these things now. But I do wonder if there's an opportunity to push back on that, to reconsider that. And if we're really stuck with this for now on. But anyway, I hope you enjoy this conversation with Liz. It was great to have her back on the show after all of this time. If you do enjoy it, make sure to leave a five-star review on your podcast platform of choice. You can share the show on social media or with any friends or colleagues who you think

would learn from it. And if you do want to support the work that goes into making Tech Won't Save Us every single week, you can join supporters like Edward in Ottawa, Johnny from Kitchener, and Olga in Brooklyn by going to patreon.com slash techwontsaveus where you can become a supporter as well. Thanks so much and enjoy this week's conversation. Liz, welcome back to Tech Won't Save Us. Thanks so much for having me back. As you reminded me before we were recording, it's actually been quite a while. I forget how old the

show is sometimes. And you were, you know, a guest on the show very early on. But you have this new book, Mood Machine, where you have done so much more research into this company and the broader effects of its model on artists, on how we, you know, listen to music. And I want to dig into all that with you. I think I would start by asking, how did you get into this topic? Like, why was this something that was on your radar and you thought was really important that you wanted to spend more time to dig into, you know, and to learn more about the broader consequences of. Thanks so much for the question. And yeah, you know, like you mentioned, getting ready for this interview, I went back and

I was listening to our interview from 2020. It's hard to believe that it was that long ago. And this question is one that I've been getting a lot, having a new book out. And it's interesting. I started writing about Spotify. The first article I wrote came out in 2017. I had started writing about it, the first article I wrote came out in 2017. I had started writing about it in 2016. And there are things that originally initially got me interested in writing about

the subject, but at all different points throughout the years, you know, there are different things that sort of keep you interested in the same subject matter. I'm sure you probably could relate to like being someone who writes about technology. The things that first got me interested in it aren't necessarily like the things that have kept me interested in it. I think back in 2016, 2017, there were a few different things that initially got me interested in the subject. One, I'm someone who's always had a foot in the world of journalism and also a foot in the world of independent music as a participant, having worked at venues and booked all ages shows for a long time. And I think that part of being involved in independent music kind of involves

this constant negotiation of what that even means. And there is a point in the mid 2010s when it became really clear that, you know, this idea of independent music was becoming not just about independence from the major labels, but also grappling with the technology companies that the major labels were increasingly partnered with that were sort of selling themselves to the independent music world as these democratizing forces for culture or as these sort of empowering tools. Another influence is that around 2015, I read this book by Asher Taylor called The People's Platform. I know you've had Asher on the show before. And you know, that book made a really big impact on me. Funny, you know, I think back on it, I actually just

kind of stumbled upon it in blue stockings in New York City, an independent feminist bookstore, but it made a big impact on me. And that, you know, whole book was sort of debunking this idea that the internet had really democratized culture in the way that some of us maybe hoped or thought that it would at a certain point. So I was kind of like fresh off of reading that book and having interviewed Astra about it at the time, kind of seeing the ways in which streaming services were sort of ingraining themselves in independent music scenes. And then the last thing of like the three things that I usually cite is also that as a freelance music journalist, I think at the time in the mid 2010s, I also was

just a little bit like bored with music journalism. Like when you're a freelance music writer, you do, you know, you interview bands about their albums that are coming up, you write album reviews, you at that time, that was like the, you know, height of like the clickbait era in journalism. So it's a lot of like getting asked to write like 200 word blurbs for like best of lists and stuff like that. And I do specifically remember just having this moment where I sat down and thought like, what else could I be doing within the realm of music journalism and made a list of three stories. And one of them was Spotify playlists. Luckily, like one of my best friends at the time worked in the music industry. And when I told her that I was thinking about investigating Spotify playlists, she happened to like have some intel

specifically into the ways that major labels were sort of seeding their tracks throughout the Spotify playlist ecosystem at the time and sort of helped put me on the track with the first story that I did. Not to give like too much of a whiny answer, but I'd say those are some of the things that made me initially really interested in the subject. And then I think that part of what made me continue to be interested in the subject was just kind of, you know, seeing how the conversation changed over the years and how like contributing to not just me, but there's like a group of writers who were sort of critically covering the streaming ecosystem, you know, another one being like David Turner, who was doing the newsletter Penny Fractions. And it just kind of seemed like it was

useful to musicians and people in the independent music world. Definitely, like shortly after we spoke in 2020 and in 2021, seeing the United Musicians and Allied Workers Justice at Spotify campaign, they put on these protests at Spotify offices all around the world. And I was thinking back on it. And like, even when we spoke in 2020, it would have been really hard for me to imagine that within a year, musicians would become so organized that they'd be organizing a global day of action outside of Spotify offices. So moments like that also, I think are things that kind of like kept me thinking like, okay, this is like worth continuing to pursue to try to contribute something that is useful. Yeah, that makes a ton of sense, right? It's a story that keeps evolving. There's always new aspects to it. And you can see that, you know, it has this real deep impact, not just on the way we listen to things with the people who

make this music as well, right? You know, communities that you're very close to have been very engaged in. And so I wonder, you know, if we're thinking about Spotify, I think a lot of people are familiar with what it is now, how it works. But where does this company actually come from? And what is the real story behind its founding that, you know, maybe has been obscured through public relations and, you know, other stories that have been told about it? Yeah, you know, it's super interesting. Even when I actually started working on this book, I started working on it in 2022, or I actually technically started working on the book proposal as early as 2019, but didn't really start working on the book until 2022. And at the very beginning, I wasn't even really sure if it was specifically going to be a book about Spotify. I thought maybe

it would just be a book about streaming or music under platform capitalism. But as time went on, I realized that, you know, actually what I was writing was a book about Spotify. So I figured, OK, let's just like go in with this. And that required, you know, doing a deeper dive on the history of the company than I'd really done before and going to Stockholm a few times and trying to track down, you know, not just former Spotify employees, which like luckily I was able to talk to quite a few of them, but also people who were in Stockholm at the time, people who were kind of involved in the tech scene and also other journalists and academics from Sweden. Spotify launched in some of its first markets in Europe in 2008,

but the company was created officially two years earlier in 2006 in Sweden. And it was started by these two guys, Daniel Ek and Martin Lawrenson, whose background was in the advertising industry. At that point, there had been other streaming startups there had been other companies that had been trying to sort of make streaming happen for a while the major record labels in response to the impact of Napster on the music business in the late 90s and early 2000s they had tried to start their own streaming services which is chronicled in the book people talk about them now there were these like hilarious failures because the major labels are so inept, to be quite frank, and like not really good at what they do. Those services did not take off. There were other apps that had kind

of come along, other companies that had tried to do something in streaming. And if you listen to people in the music business, you know, they'll say Spotify was the first company that really got the technology right, that created an app that like felt like the music was on your computer when you pressed play. Real emphasis on like how the user interface was more sleek and loaded more quickly than others. But something that I think is important to remember and that I try to emphasize a lot in the book is that these people who started Spotify, they didn't come from the music world. They came from the world of advertising. And I think that something else that's important to remember is that at that time in the mid 2000s in Sweden, piracy and file sharing at that time was popular everywhere, but it was really popular in Sweden in a different way. It was politicized in a different way. It had a different sort of culture to it.

There was more of a question, you know, where maybe in the United States it was, you know, more widely deemed to be something like bad that needed to be stopped. In Sweden, there were these bigger cultural questions like maybe file sharing actually is a good thing. You know, maybe the music business and the entertainment industries have too much power and maybe like something does have to be done to make art and culture and music and movies more accessible. There was even a pirate party in Sweden. There was a political party that was completely founded on being pro-pir saw as their competition wasn't other apps in the music business, but their competition was the Pirate Bay because it was so popular and it was what everyone used in Sweden to listen to music. Spotify sort of took on this interesting character in Sweden where on one hand, they were sort of positioning themselves as this legal

alternative to piracy at a moment where not just people in the music business, but also lawmakers because of how politicized piracy was there. They were like really looking for something that would solve this. You know, I went and spoke with in Sweden, this person named Rasmus Fleischer, who is a co-author of this book called Spotify Teardown, but he also was a founding member of this organization called Pirat Baran, which means the Bureau of Piracy. And that was formed in response to this Hollywood lobbying group called Anti-Pirat Baran, which means the Anti-Piracy Bureau. And in response to Anti-Pirat Baran, this group of musicians and like punks and ravers and hackers, they were like, okay, well, if you're

anti-Pirat Baran, we're Pirat Baran. And actually we're a lobbying group too, but we're lobbying, you know, for the rights of people who think file sharing is important. We're a pro-piracy lobbying group. And I think that kind of speaks to like the conversation at the time, or as, you know, some of those folks told me, some of the cultural climate that someone like Daniel Ack or Martin Lawrenson might've been capitalizing on a little bit. Their initial idea for Spotify was, you know, to try to create a content delivery mechanism that provided sort of the ease and experience of piracy, but that was funded by advertising. So at the very beginning, there wasn't any idea of like a subscription model. They just basically wanted you to feel

like you had your piracy enabled collection on your hard drive, but it was all funded by ads because their background was in the advertising business. It's such a fascinating story. And so much of it, I didn't really understand, you know, like I had a basic idea of where Spotify came from. But, you know, this notion of like the climate in Sweden at the time it was emerging was something that was completely new to me, you know, and the fact that these streaming services, there were other versions of them in the past. And as you describe in the book, those streaming services that were launched by the labels

were described by some people as a huge conspiracy, you know, not meaning conspiracy theory, but, you know, like all these companies coming together to like capture this kind of market. And, you know, ultimately, even though Sweden has this particular kind of political climate around piracy at the time, what ends up coming out of it is this kind of like ultra commercial version of what this should be. So what is the Spotify model that eventually emerges and how does that evolve as Spotify has to, you know, kind of go into negotiations with these major labels to ensure that all the music that they own and control can be on a service like that. So like I mentioned, the model initially was to create a product that would offer, you know, this large library of content. It actually is painful to me to use the word content, but there's kind of no way of telling the history of the company without using some of the language of how they talk about music. But every time I say the word,

I'm like, Oh, I can't even, it's hard to take myself seriously. But their initial model was, yeah, this product that would give you this large library of like, content, you know, any music you could possibly imagine supported by ads. But you know, if you listen to these early conference appearances by top executives, early employees, you know, it's quite clear that the beginning, it wasn't even really sure if they would do music. They had also considered doing a video streaming service and their initial patents don't specifically mention music. It's just kind of like any type of media, really. And eventually they ended up landing on music. In some of the interviews, they talk about how the small file size was part of that. Probably also everything I just explained about

the weak status of the music business and things like that. When I was reading that section, it reminded me a bit of like Amazon, you know, and how Bezos like started this business. And he was like, okay, we'll do books because it's easy to do books. It's easy to ship them. Like there are all these benefits of doing so. And then we can expand into other things. Like it really spoke to me when I was reading what you wrote about Spotify that it was like, oh, they chose music because it seemed like it made sense and was easiest to do not because they were particularly interested or cared about music. Yeah, absolutely. And I mean, now you see Spotify moving on to like podcasts and audiobooks. And in

some ways, it's sort of illuminating how it never really was like, you know, about music per se. In some ways, music just sort of allowed them to grow this user base and grow this platform pretty similar to Amazon and books for sure. But then, you know, it was really the negotiation with the major labels through which the subscription tier emerged as well. So currently, you know, Spotify makes most of its money through subscriptions. So people paying, you know, 10.99 a month or as we just were discussing more in a place like Canada, because Canada has rightfully imposed taxes on streaming services in order to fund local music and content, other content, as they say. But yeah, you know, most of their revenue comes from subscription services. Advertising is still part of it as well. And then increasingly, a service like Spotify,

they refer to it as a two-sided marketplace when they speak to investors about what their model is, where on one side, they're selling a product to the listener. And on the other side, they're selling a product to musicians. So they also increasingly, they refer to it as quote unquote marketplace, the side of the company where they're selling advertisements to musicians and promotional opportunities like algorithmic promotion in exchange for reduced royalty rates or banner ads or like shelves on the homepage. So like that also is part of their business model as well. But the vast majority of the revenue comes from subscriptions.

I think that makes a ton of sense, right? And I think it's interesting to see that evolution and the influence that the labels have had through that negotiation, through that leverage that they have, where even though they don't own this streaming service like the one that they used to, they know, they have still influenced how it has evolved over time and the types of model and the way that this model works and the way that they

benefit from it. Right. I feel like one of the pieces that really stood out as I was reading it was how much the labels have been able to, and the major labels in particular, like, right, like the big three labels, Warner, Universal, and Sony, have really been able to profit immensely from this model, from Spotify, even as we have these discussions about how musicians aren't making, or the vast majority of musicians aren't making nearly enough and we're not paid enough through this model. Can you talk about that distinction and how that actually works? So part of it is in order for a service like this to function, the streaming services have to sign deals with the rights holders. So streaming services don't sign deals directly with musicians.

They only sign them with rights holders. So the rights holders that they'll sign deals with are Sony, Universal, Warner. For the independent labels, it's this industry trade organization called Merloon that sort of gives the independent labels a collective voice in the negotiating processes that has existed since it formed just before Spotify launched in 2008, actually. And then they'll also sign deals with distribution companies. So these are companies like maybe you heard of DistroKid or TuneCore. There's these sort of like self-serve companies that like any, you know, if you like recorded a song on your computer today and were like, I want to put this on Spotify, you could go set up an account with DistroKid and distribute your music to streaming services in a few days. So the streaming services maintain deals with

like these, these rights holders. And because major labels, you know, sit on the rights, the copyrights of so much of the back catalog of what we think of when we think of the history of recorded music, they own a huge percentage of the recorded music market. That puts them in this really incredibly outsized position when it comes to negotiating each of those deals. So each rights holder has its own deal with Spotify and those deals dictate the terms of royalty rate, the terms under which the relationship exists. So those contracts are incredibly secret, you know, like even artists that are on those labels, like artists on major labels, artists on independent labels that work with Merlin, like don't get to review the contracts. It's only a very like small privileged group of people who ever even get access to seeing what's in those contracts. But yeah, the majors, you know, they have this like extreme outsize influence and negotiating power when it comes to working out those contracts because of how much music they own the rights to. So in the initial deals with Spotify and other streaming services, the major labels were able to negotiate for themselves things like really big advances, equity stakes in the company. They were able to have certain perks in their contracts

having to do with marketing and advertising. And they also were able to negotiate for themselves specific minimum rates. So on the free tier, for example, they might have a per user rate or a per stream rate that they get that is not necessarily guaranteed to other rights holders because of the pro rata model that dictates the entire royalty system. A lot of how I write about it in the book is still drawing from, you know, there's this one contract that leaked and there's a Sony's contract with Spotify in 2015. And it's still largely kind of like the best glimpse that you can kind of have into what one of those original initial contracts with a streaming service at the

time would have looked like. But there are certain things about it that were kind of terms that had come to be standard at that time anyway. Like it's really interesting to see that, but it's also really concerning, right? To know the outsized power that they have in shaping how all this works and, you know, how that then affects everyone else. And I think we're going to come back to this to talk about it a bit further, but you mentioned earlier how, you know, playlist was one of the ways that you really got into looking into Spotify as a company and how that works. And I feel like the book deals a lot with that, right. As an important aspect of how Spotify as a company works and how it has evolved over time to be so focused on these playlists, on these particular playlists around moods rather than just genres,

and how originally starting with human curation that has moved much more into the use of algorithms and AI and these sorts of things to shape what is going on here. So how did Spotify move to become so dependent on playlists? What was the incentive there? And how have we seen that change the way that people consume music and, you know, just shaping the way that people use this platform more broadly? Yeah, it's a great question because looking back at the early history of the company, in the early days, the experience of using Spotify was much more like a search bar. You would have to know what song you were looking for, what artist or what album you were looking for. the United States and after they launched in the United States when you know there started to be much more of an emphasis on like how do we grow beyond just the dedicated music enthusiast type

listeners and how do we get this more like mass generalized audience there's like a lot of things that impacted that like you know one was like launching the United States where there's more competition there were more pre-existing music services. This would have been, yeah, like 2011. Something else also is that compared to, for example, Sweden or some of the other European places in the United States, like the iTunes store was really popular in a way that it didn't catch on. And certainly in Sweden, maybe other places. So they were sort of like competing against that competing with pandora and then additionally it was also around the time when you know they started taking vc investment like before they even launched in 2008 but at that point they had you know taken on more rounds of vc investment and i think it was

around the sort of time where like maybe the creeping pressures of that had started to kind of like you know catch up with them like it was definitely a moment where there is a strong need to grow the user base and to figure out new ways of convincing people to subscribe or to get onto the platform. So I think that the shift into curation is very much like a growth tactic. And something else that was super interesting was in my interviews with former employees, people who were close to the company at the time, like learning that there's actually like a specific moment where the company, after they launched in the United States, had actually like hired a research agency to conduct research on their own user base

and found that, you know, this would have been like as they were starting to experiment with playlists and more curation and found that the vast majority of listeners were coming to the platform for more of what they would refer to as like a lean back experience than a lean in experience. So it seems like both because they were trying to grow, maybe as a result of some of this research that they'd been doing on their own user base, they really like ran with that and started to really optimize for more of this lean back listener, maybe this more like, you know, user who isn't super concerned with the artist or the album, or even really, really what they're listening to. And it's happy to just like, you know, pick up playlist category and hit play. I found that so interesting, right? You know,

the discussion of this type of listening experience and like, you know, we have had that, I guess, to some degree with like radio for a long time where you can just like turn this on and you can listen to it and you don't need to worry about like picking a track or, you know, picking what vinyl record or CD you want to listen to or what have you. But it also resonated with me because I think about how, you know, like a decade or so ago, I feel like I used to be much more active in, you know, trying to find new music that I enjoyed. And it was right around the time of like the move to streaming that I started to find that like a lot more difficult to do or I found it hard to do like with the streaming model but that did also correspond with like my life just getting busier and that being one of the things I felt like maybe I didn't have enough time for but maybe also because the mechanism for

how it worked or how people were listening to music changed that it felt more difficult to do that so so I you know that commentary is I think my question on that is like, do you think that this lean back experience is something that people are already looking for? Or do you think that the Spotify model like encouraged people to see that as the way that you listen much more often? Yeah, I think that something that is helpful to, you know, keep in mind is the reasons why a streaming service would be incentivized to prioritize lean back listening and why they would be incentivized to maybe make that be the first thing that you see when you open the app be like a series of playlist categories or you know really convenient mixes optimized towards background listening. Part of that is

because this is actually something we talked about in 2020 also, but streaming services like any other platforms are trying to boost engagement on their platforms. So playlists are a tool of boosting engagement. You know, in the early years, these were playlists that were made by a team of in-house curators. There still are a small number of in-house curators employed in each market by Spotify, but over the years has grown much more personalized, much more based on showing you what is in your listening history, much more focused on maybe the novelty of something like an AI playlist or a day list. And the goal is, you know, engagement goal is like risk management. They don't want to risk

losing you as a subscriber. So they're basically just trying to show you whatever is statistically the thing that you are most likely to click on and keep streaming because success is something that is gleaned from the streams continuing to flow. Leanback listening has always existed. The whole history of radio is also in some ways a history of leanback listening. I would argue maybe that changes when you start looking at things like community radio or internet radio. A great example being, you know, a radio station that I really like WFMU. It's a like free form community radio station in New York city with a really, really active like chat feed. And there actually are a ton of radio stations, online radio stations that also have these really active chat feeds that are sort of, I would argue like a type of radio stations, online radio stations that also have these really active chat feeds that

are sort of, I would argue, like a type of digital public space where people are like listening to music and talking about it at the same time. And I just say that to kind of push back on, you know, people always are like, oh, lean back listening has always existed. Like, you know, look at the radio. But it's really like, again, like, you know, we're always talking about not just technologies, but like the WFMU like constantly active like chat space on their website like this clearly is not just a lean back experience for like most of the dedicated listeners of this station sorry a little bit of a tangent but certainly streaming did not like invent the concept of the lean back listener but I do think that there are a lot of reasons why over the years, Spotify and other streaming services have been very incentivized to make this the mode of listening that is most championed on their interface because it's the thing that's the most profitable for them. with like the shift toward making playlists around moods and playlists that are algorithmically curated to ensure that people are constantly having these like updated feeds that feel

personalized to you, you know, regardless of how accurate that data is or what they're actually doing behind the scenes. Can you talk about that move? And in particular, you know, you mentioned there how this is more profitable, how the company then starts to target music, licensed music, that it's going to be able to put in these types of playlists much more cheaply to improve its bottom line. of like about the concept of the chill playlist or the chill vibes playlist as this streaming era phenomenon and trying to sort of historicize it and think about maybe why chill has sort of like you know emerged in the streaming era as the kind of like millennial version of easy listening which is a type of radio that has existed for a very long time. And it's interesting looking into like not just the broader history of mood music, the broader history of like new technology formats trying to sell themselves to users through the use of music being sold as a tool of mood

stabilization. I sort of historicize it or contextualize it as just a tactic of advertising, you know, like advertisers and marketers, you know, it's like a classic tactic of advertising playbook to cater to people's emotions, to cater to people's moods. So it kind of, in some ways it makes sense that like tech companies that don't know anything about music that aren't interested in hiring people who know things about music, you know, who really aren't interested in teaching you about music at all, who really just don't care about music would be seeking out more methods of engagement, like outside of that. A quote from the book that actually like really stuck with me was this one former Spotify employee. I was trying to understand

like, why did chill playlists become so popular? Like, you know, I've been writing about this phenomenon of the chill playlist for so long, but I still don't understand in some ways, like, why is it so popular? And this person was just like, we made chill playlists because people consumed them. And if you're a playlist curator, and you like have a new job at Spotify, you just want to kind of make your metrics go up, you want to show that you're doing a good job. So you're going to make playlists that are that are popular, and show playlists were always popular. So we always just made them. In some ways, it's like, you know, this incredibly boring reality that so many people who work at tech companies

are just tasked with moving a metric. And it's really like, you know, that's all that it is. And I feel like that's something that came up both in conversations with editorial curators, something that came up talking to a lot of former machine learning engineers, like, you know, it came up talking to people from across the companies, you know, they would say things like, I know it might seem like these are kind of like nefarious decisions that are being made, or there might seem to be like something like kind of like, you know, grim, but a lot of times it's just someone who has a job and their job is just to make this number increase. And they're just like changing

things around on the interface. They're just curating things in certain ways to just make this number go up. It speaks to kind of like what happens when the only value is growth. And that's like the only thing that people are being tasked to care about. Yeah. It's so depressing to read about that. And like, you have so many more examples of it in the book, but as you mentioned, you were talking about, you know, something that really resonated with you in that part of it. And there was a quote that I wrote down, the conquest of chill reflects an industry content to profit from a world of disconnection. I feel like that would resonate a lot with the people who listen to the show and the people who are concerned about, you know, the broader impacts of what these services are having on the ways that

we consume culture and the ways that we think about culture. Have you seen that shift in the era of generative AI as well? These playlists feel like as they have become more reliant on licensed content and on algorithmic feeds that they seem set up for the use of something like generative AI to keep finding the next stage of this model. What have you seen over the past couple of years? I mean, I think that all of the concerns around generative AI content, displacing the work of real working musicians are really important and valid. You know, like as recently as last week, we're just hearing the most like abysmal, depressing types of sentiments coming out of the CEOs of generative AI music companies. I'm not sure if you saw this quote floating around.

It was the CEO of Suno, which is kind of one of the big generative AI music or audio companies. And he's on this podcast saying, like, most people do not like making music. Like, you have to learn instruments. You have to get good at instruments like you have to learn instruments you have to get good at instruments you have to learn how to use production software it's really hard and it's just like it's the most like out of touch it's like yes like learning to play music does require some effort it requires like having to want to learn how to play music or you don't learn and you make music and with a few simple chords. And that's what's beautiful about like folk music and punk and like minimal music. And like, you know, like it's just with the generative AI space, so many reasons to, you know, firmly believe that these people are like absolute villains of culture and music and creativity and community. I think for me, and like the thing that I'm maybe equally concerned,

if not like more concerned about is in addition to generative AI content, which kind of tends to get a lot of attention, a lot of hype around it is just the ways in which automated systems and machine learning and algorithmic recommendation over the past 15 plus years have shaped not just the music and the sounds of music, but like how we understand music and how we relate to music and how we like form meaning around music and our concept of what music is for in our lives. So, you know, thinking about the ways in which algorithmic recommendations have come to increasingly influence discovery feeds or recommendation feeds, how algorithmic decision makings have come to shape the types of words and phrases and terms that get used to describe music within the interface of a company like Spotify and just like the kinds of connections that get made or don't made. There's a few examples that are instructive, which is I think right now, if you have a Spotify account and you open Spotify, chances are like

what you're going to see when you first open the feed is something called daily mix, which is basically, you know, Spotify has a taste profile on you, which is datafied version of your music listening history, maybe sorted into different buckets based on different types of stuff that you might listen to. When you open the feed, you're getting these daily mixes. And it basically looks like just a bunch of mixes based on stuff you've listened to in the past. And I think that even people I know who are discerning music fans who would never listen to generative music, who are very like curious listeners too, could easily fall into just the habit of they open this app, they think, oh, well,

this is basically stuff I listened to already. So like, I'm just gonna listen to this because I recognize most of these artists and like, yeah, sure, like whatever, like hit play. In doing so, it's kind of like, I think, a slippery slope into developing listening behaviors around a recommendation pipeline that is controlled by discovery mode, which is where, you know, musicians and labels can accept a lower royalty rate in exchange for algorithmic promotion. And so when you're listening to a daily mix, like a lot of the music that you're being recommended is actually being recommended to you based on a pre-existing commercial deal between Spotify and various rights holders that is, you know, totally unknown to you. And because of that, there's also a lot of music that you're not being recommended because it's

people that are not participating in this payola-like scheme, not actually payola, but it has some similarities with payola. Another example of something that like, you know, influenced by tools of algorithmic recommendation and machine learning and AI that you might see when you open the platform might be something like AI DJ, which is something that they unveiled last year, or sorry, at this point is probably two years ago. AI DJ is sort of like partly the culmination of something that Spotify has been after for a while, which, you know, if you look back over the past decade, interviews with executives for a really long time, they've been after this thing where they want someone to be able to open the app and just hit a play button. This is something that like,

you know, in 2018, when Spotify went public on the stock exchange, their co-president, Gustav Soderström, talked about how Spotify's goal was self-driving music, you know, or like, and even earlier, one of their first senior playlist executives talked about how the goal was to just open the app and have one button that you press. So AI DJ is kind of something that they rolled out where, yeah, you know, it's basically just a play button and it plays you different slices of different types of algorithmic recommendations. So there's not one, the algorithm, there's all of these different pools of data. There's all of these different models that are built using these pools of data.

And then there's different products that draw from these different recommendation models. So AIDJ kind of like draws from these different recommendation models. So AI DJ kind of like draws from these different recommendation models and it might be like, you know, all right, like coming up next is some music you listened to in 2023 or coming up next is some music from your release radar, which is like, you know, new music from artists that you follow.

And all right, now you're going to listen to, we're going to listen to a block of songs that are metropopulous, which is like a genre that they made up stuff like that so it's like ends of the voice that's in between the blocks of songs is a generative ai voice trained on one of their employees i think the thing with ai dj that gets at this thing that you were asking about about this kind of like alienated isolated like atomized way of relating to culture is that it's such a good juxtaposition

again with like, you know, what the experience of listening to like an actual DJ might be. You know, when you listen to the AI DJ, the voice in between the songs is never telling you, this is the artist that is about to come up. This is the album where it's from. This is where this person's from. This is when they made this album. And you know, this is the label that it's on. It's just here's some songs that you really liked in 2022. You're really going to see yourself in these or it'll be like, here's some songs from your like, you know, indie chill playlist or you know, like, it's all about like how it fits into your taste profile. It's not at all about pointing you out towards the world of music and culture and trying to make connections and say like, all right, this next song came out in, you know, 1993 and it was

on this label and it was made by these two people working together. It's nothing actually about culture or trying to fulfill, you know, the role that a DJ might historically play that is more, you know, maybe educational or trying to actually kind of like introduce you to something. So, you know, just like a couple examples of things that I think are examples of the ways in which the increasing algorithmic nature, hyper-personalized nature, AI driven nature, you know, if you listen to interviews now, like even since I finished my book with Gustav Soderstrom, who like, you know, everyone's emphasis is always on Daniel Ek when they talk about Spotify, but there actually are other high-level executives like Soderstrom.

There's so many podcast interviews with him. Like he does so much media and like is actually like another really interesting glimpse into the state of the company. But he'll say things like, you know, when it comes to Spotify now, like the AI is the product. And like, you know, meaning that the interface, when you open the app, you know, the ability to like have your whole listening profile and almost like in a concierge type way, like, you know, be recommending you like the perfect thing at the perfect moment, whether it's like a podcast or an audio book or your stuff in your listening history.

Like, you know, at least in this executive's perspective, he talks about that being the thing that they're selling. And I think that that has a lot of consequences for people's understanding of music. Yeah, that makes a lot of sense. And I like your reframing of that. So it's not just about generative AI and what has been happening for the past couple of years. But, you know, this broader use over the past 10 or 15 years of machine learning, artificial intelligence, all this kind of stuff to really reshape not just how we interact with music, but so many other things in society as these companies have been

trying to do that. So I really take your point on that. I was wondering, you know, when it comes to the artists themselves, you know, we touched on this earlier, like, what is the effect of this? Because one of the things that you talked about in the book, one of the things I remember, you know, reading about in the media was this change in the streaming model that was pushed by Universal last year, and you know, the wider effects of that. But meanwhile, on the other hand, how a lot of artists were pushing for a change in the streaming and royalty model as well, but they were looking for something very different, right? The change to what you call a user-centric royalty model.

Can you talk about these two competing proposals and what the effect of the newest direction of streaming has on artists and people who make music? So for years, musicians and advocacy groups and musicians unions have been putting forth a proposal for something called user-centric streaming. So currently streaming royalties are dictated by this incredibly complicated and convoluted system called ProRata, which is basically this like really complicated revenue share system. But the basics of it is that, you know, payments are based on market share, or as they call it stream share. So if you're a rights holder, you're a record label, and your catalog amounts to, you know, maybe like 3% of all of the streams

that happen on Spotify in a given royalty period, then you would be entitled to 3% of the eligible royalty pool per the terms of your contract with Spotify or whoever represents you. So there's so many things that are complicated about this. One being that what the eligible royalty pool is, is different depending on your contracts with Spotify. And it's impossible to know what everyone's contracts are because they're all completely secretive. But the basics that you hear from the music business are that 52% of the revenue goes into this pot that goes to recorded rights holders. And then it's pro rata cut based on stream share. The model that artists and unions and advocacy groups have put together is called user-centric, where instead of this really complicated system that is so

confusing that most musicians actually struggle to understand, and most listeners, if you ask, probably would have no idea what any of this is. User-centric instead would just say, I am a Spotify user. I pay $10 a month. Spotify gets a 30% cut, which like, you know, they have since the beginning. And then the other 70% goes to the artists that I stream. So like if I spend, you know, my whole month only listening to the band Lorraine, then Lorraine should get my $7. That would be much more straightforward for musicians and users to understand, I think. And it actually, you know, I have complicated feelings as expressed in the book about the music industry's obsession with fraud and scammers. But it also would help to address

some of those issues too, because if I set up an account and like from my account, I'm just like, you know, leaving some AI music on repeat, like all night trying to like scam royalties for music made by AI bots, like your $7 only goes so far, regardless of how much you stream. So it would kind of help address some of the stuff with fraud also. So there's a lot of reasons why this model makes sense. But instead, the major labels put forth a different series of proposals, you know, supposedly aimed at like stomping out white noise bots. And as they very disrespectfully to most musicians also say that they're not just stamping out scammers and bots, but also amateurs and hobbyists, which they're happy to sort of throw under the bus with like all of these AI fraudsters, which I think is highly problematic to put all of those different types of musicians in the same

pool as a problem. You know, in some ways, I feel like the music industry has really kind of like waged a campaign against what they see as like deprofessionalized musicians. Lots of complicating things to unpack there. You know, even artists or musicians who are making music. You know, a lot of musicians on one hand are pursuing professional careers and would like to consider themselves professionals, but because of how stacked against them the status quo of the music business is, they're unable to reach the threshold to be considered, quote unquote, serious musicians in the system, and then to kind of like throw them in with the bots and the scammers. I guess I should explain the Universal music group proposed system suggests that you know any artist or any track that generates less than a thousand streams per year just wouldn't get paid at all and then there are also like all types of other music that would be demonetized so like stuff that is considered

non-music noise content so like field recordings recordings, ambient rain sounds, like white noise, you know, some of this like fair enough, like there are a lot of people out there that are like trying to kind of just like scam the system with like AI generated rain sounds and like just to try to juice as many royalties as possible. And it makes sense that they would be trying to seek a solution to that. But one other thing that I raise in the book is that it also, you know, has consequences for artists that make music with field recordings. Like I interviewed the director of Smithsonian Folkways, which is a legendary record label in the history of American music that has, you know, existed for decades. decades that considers itself to be stewards of an archive of a lot of really like historically important field recordings and nature recordings. And from their perspective, these are types of recordings that they never intended to get rich from or that they never intended to be highly

profitable undertakings, but that the positioning of this material as, you know, spam is a disservice to the work that they try to do as responsible stewards of the archive and also could have impacts on the ability of users to discover the material within like the streaming landscape if it's being like, you know, deprioritized or considered, yeah, like spam content that has to be dealt with. In addition to all of the things about the model that independent musicians have concerns with, it's also just that the fact that the model was, you know, created by Universal Music Group, advanced by Universal Music Group, Universal Music Group piloted it with Deezer first, which is a French streaming service. And then according to reports in the Financial Times, when they renegotiated their contracts with Spotify in the season

following that, they made it a requirement that Spotify adopt what they call artist-centric streaming. I think it just further proves the influence that a company like Universal Music Group continues to have, not just over their relationship with Spotify, but over the entire streaming model beyond even just Spotify, you know, it's very risky when we start letting companies like Universal Music Group and these big tech companies be determining like who is a serious artist and who is an unprofessional hobbyist who doesn't deserve to make like any royalties at all for their work. These are really slippery slopes, I think. Yeah, I can definitely see where the frustration lies in that, right? You know, especially if you're an independent musician, you know, making music that you care about and a company like Universal or Sony or Warner has

such a huge effect on how you are going to get that music to people, regardless of whether you're dealing with them at all. But, you know, we've talked about a lot of the issues with the streaming model and with Spotify in particular throughout this conversation. And so I wanted to end off by asking you, because at the end of the book, you have a number of initiatives that people are trying to take to, you know, push back against this model to try to envision something else, not just to improve the streaming model by increasing streaming rates or moving to a user centric royalty model, but also thinking about how our relationship with music can be different. And I wonder what you think is, you know, the best hope in some of those initiatives that you outline

that really resonate with you. Yeah, for sure. I mean, like I mentioned earlier, like I definitely think that the kind of emerging new music labor movement, which I write about towards the end of the book, like in some ways I find to be really inspiring for a lot of reasons, not only because it's so powerful to see, you know, when I first started writing about Spotify and streaming, when my a somewhat like, you know, meaningful royalty check. There's a lot of sort of like fear of backlash from the services. And I would have artists like DMing me saying like, thanks for saying the things that none of us feel like we can say and stuff like that. It would have been, you know, hard to imagine how much things have now changed and to see how much it has changed and it would have been, you know, hard to imagine how much things have now changed

and to see how much it has changed and that there have been, I think there's just a really strong instance of like, I think that's what's happened is that a lot of musicians have like looked at the penny fractions that they're paid by streaming services and realize that the power of a collective voice is worth more than like the $2 royalty check that they might be getting per royalty period. And that has been really, in my opinion, like really inspiring. And like, it's a moment where the music business has really pushed this idea of the musician influencer or the atomized content creator as the sort of model independent musician, you know, we didn't really get into this as much. But there's a lot of ways in which I think the royalty model has really like incentivized solo artists.

Daniel Ek talks about like the model musician as this kind of like creative entrepreneur artists that they sort of like put forth in their marketing and in their sort of like Spotify for artists campaigns and things like that. So at a moment when this sort of solo creative entrepreneur has been upheld as this kind of like model musician, I think it's even more powerful that so many musicians are sort of like pushing back against that by forming not technically unions, but these solidarity organizations. So in the United States, it's considered illegal for independent contractors to unionize or do boycotts and musicians that fall into that category. So for years, you know, musicians have been told like, you can't collectively organize, you can't form unions, it's illegal, you'll be deemed an illegal cartel. And there's actually a piece of legislation called the Protect Working Musicians Act that specifically, and specifically this piece of legislation that they've been working on the past couple of years with Congresswoman Rashida Tlaib called the Living Wage for Musicians Act. And it's based on some pre-existing laws that already exist around digital radio. It's rooted in a lot of research. It's rooted in examples of how something like this could function. And it seeks to add an additional royalty on top of the pre itself, but because of the

organizing work that it's really championed and the way in which I think it has kind of like helped musicians in some instances, like see the power of collective organizing, maybe see themselves as workers or feel more comfortable, like talking about what they do that way. What you talk about there reminds me a lot about how Amazon used to treat authors as well, right? You know, kind of promoting this idea of a different kind of self-published author that is, you know, releasing books every like three months because that's what's going to work best for the algorithm. Like another example of these platforms, you know, and using their forms of like algorithmic

management, for lack of a better word, to like try to change how creative people or artists, you know, see the work that they do, whether it's in music or, you know, in writing books and publishing or what have you, right? I'm sure that they've been doing it in the other sectors too. And it's always fantastic to see when these people are coming together to like push back against a model like that. But one of the other things that you talked about near the end of the book, and you know, I'm a huge fan of libraries and think that they're just an essential piece of social infrastructure in our societies that feel like an example of like a bygone era. Like you can hardly imagine our current political systems coming up with something like a library and funding it.

But because they're there, you know, we still have them, we need to defend them, we need to expand them to make sure that something like that can grow. And you talked about how libraries can also be this important piece of thinking about how we interact with local music in particular in a very different way. Can you talk to us a bit about that to close off our conversation? For sure. So after my part of my book about the new music labor movement, in my conclusion, I sort of outlined a constellation of different ways in which perhaps we could collectively imagine revaluing music. In the streaming era, we talk so much about how music has been devalued, both financially, but then also in terms of the

watering down of the relationship between listener and musician that has happened because of all this lean back listening and like algorithmic homogenization. So what would it look like to revalue music? So I talk a bit about like different musicians working on like cooperative streaming services and different like things like that. But also something that to me is like extremely inspiring, not because it just represents like a completely different way about thinking of the value of music in society, but also because it's already happening is, yeah, these examples of around the United States and Canada, local public libraries, you know, local public libraries already offer in many cases, subscriptions to cardholders for things like Hoopla, which is like basically like, you know, Spotify for the library.

It's, you know, you'd be surprised how many like major releases you could find on something like this. It's basically like a library version of Spotify or Apple Music or something. But outside of those types of resources that many public libraries offer to cardholders, there's also these instances of libraries launching these local music streaming projects. So these really small scale local music digital libraries where they'll get together a group of people who are really invested in the local music scene to act as curators. And then local musicians can send in their records.

The local curators will like each round pick a selection of local albums, and then the library will license the music directly from the artists for a set fee for a couple of years in order to create these small scale, but I think like really meaningful collections of digital music that anyone can stream library card holders can download. And it also to me is interesting because even though these license fees in many cases, like aren't a ton of money, you know, it's usually like two or $300 to license the record for a ton of money, you know, it's usually like two or $300 to license the record for a couple of years, you know, you'd be surprised like how much more impactful that could be than the number of streaming royalties an artist might see from a streaming service in

a similar number of times. It's not like changing anyone's life, but it could be, you know, enough to like get a band in a studio, like recording their next like demo or single, or it might be really helpful and like pressing a run of cassettes, you know, like these types of resources, like are not completely meaningless. Even if it might seem like a really small amount of money, especially when you're talking about small scale cultural production. And something else about these projects that I think is really interesting, too, is it actually is a library in Canada in Edmonton. I interviewed one of the librarians from the Edmonton Public Library who have been really like instrumental in sort of like setting this context within which a lot of other libraries have started viewing their their projects, which is like to not think of these just as digital music collections, but as digital public spaces. And to think like, you know, not just about this repository of music, but also like all of the connections that get built around this repository of music as well. The library

streaming projects are really interesting to me for a bunch of reasons. One of them is because, like mentioned earlier, and as I'm sure so many people who come on your show say, when you're criticizing technology or when you're writing about technology, you're rarely ever actually talking about the technology itself. And oftentimes what's being critiqued is power or what's being critiqued is capitalism or industry or corporate consolidation or or just really, you know, putting profit over people, you know, like these are the things that you're criticizing. And with the library streaming programs, you know, it's a great example. Like these are still instances of a collection of music being held on a central server, it being presented to people through an interface that looks pretty similar, like Bandcamp or a streaming service. And then, you know, using the internet to stream the music on your computer.

Like in a lot of ways, the technology is actually kind of similar to these other streaming services that we know, but because they are being built in a lot of times alongside the communities that they're meant to serve, taking in feedback from musicians. It's a great example of like how, you know, these technologies that we rely on could look like so much different if they're actually built with the people who rely on them in mind, or even, you know, as partners working together. Yeah, I love that so much. This is an example of how something, you know, that we interact with every day can work very differently when organized around a different set of values. So I think

that's fantastic. You know, there's so much more we could have dug into, but Mood Machine is a great book. I highly recommend people pick it up. Thanks so much for coming back on the show, Liz. Thank you so much for having me. Liz Pelley is a music journalist and the author of Mood Machine. Tech Won't Save Us is made in partnership with The Nation magazine and is hosted by me, Paris Marks. Production is by Eric Wickham. Tech Won't Save Us relies on the support of listeners like you to keep providing critical perspectives on the tech industry. You can join hundreds of other supporters by going to patreon.com slash tech won't save us and making a pledge of your own. Thanks for listening and make sure to come back next week.