Eric Drott, "Streaming Music, Streaming Capital" (Duke UP, 2024)
01:34:12 Share

This episode is brought to you by Disney's Mufasa, The Lion King. Get tickets now for the ultimate family holiday movie experience. Reunite with the characters you know and the untold story you'd never expect. Witness Mufasa's rise from orphan to king and see how the legendary villain Scar got his name. Disney's Mufasa, The Lion King. In theaters everywhere this Friday. The Kingdom Awaits.

This episode is brought to you by Amazon Prime. There's nothing sweeter than bacon cookies during the holidays. With Prime, I get all my ingredients delivered right to my door, fast and free. No last minute store trips needed. And of course, I blast my favorite holiday playlist on Amazon Music. It's the ultimate soundtrack for creating unforgettable memories. From streaming to shopping, it's on Prime. Visit Amazon.com slash Prime to get more out of whatever you're into.

This episode is brought to you by Allstate. Some people just know they could save hundreds on car insurance by checking Allstate first. Like you know to check the date of the big game first,

Welcome to the new books network.

Welcome back to New Books and Music, a podcast from New Books Network. I'm your host, Gummo O'Claire, and today I'm joined by Eric Drott, who's Professor of Theory at the Butler School of Music at the University of Texas at Austin. We're going to be talking about his book, Streaming Music, Streaming Capital, which was published by JIC University Press this year, 2024. Welcome to the show, Eric. Oh, thanks for having me, Gummo. So just to start off, I wondered if you could just tell us a little bit about your kind of academic and maybe professional background, and then what has drawn you to writing this book?

Well, I'm a professor of music theory at the University of Texas at Austin. I've been here for about 20 years.

Before that, I was at Yale, where I taught for a few years after having graduated there, got my PhD there back in 2001. Yeah, my background's actually in avant-garde music and sort of experimental music traditions. That's what I wrote my dissertation on. That's what, you know, early on in my career, I worked on a lot.

You know, it's a kind of circuitous route to get to this book. I'll keep it brief. But basically, I wrote my first book, Music in the Old Revolution, which was

about how the sort of political events of May 68 in France, you know, resonated throughout different music cultures in France. One of the things about that book is I was very attentive to the ways in which genre mediates the politicization of music. That then led me to think more deeply about genre. I wrote a big article on genre that got published a couple of years after that.

And then I was continuing doing work on genre subsequent to that. And one of the things I kind of stumbled upon almost by accident was, you know, doing a Google scholar search, you know, as many academics will do, you know, just see like, is there any literature on this?

music genre that I'm missing. And one of the things that struck me was all these article titles were popping up in my Google scholar search, which were more or less illegible to me as a music scholar. These were, you know, basically articles coming out of music information retrieval that were mainly concerned with automated systems for genre recognition that were being developed by

Partly as a sort of research activity, but partly as a kind of industry application for digital music platforms. And this got me interested, okay, well, how are they defining genre? How is it similar to, but how does it differ from music?

both, you know, kind of musicological understandings of what music genres are, but also I think everyday kinds of practical understandings that listeners and music industry personnel and musicians themselves have of the genre. And then from there, it was kind of natural, right? I mean, basically the question that's led me to wonder is, well, who is the audience for whom these...

Articles are ultimately being written. These algorithms are being designed. Who's the sort of customer base for this? And the obvious answer is, you know, all these streaming platforms, which were, this is about 2014 or so, that were, you know, really kind of hitting their stride at this point. I mean, Spotify had been around for about six years.

By that point, it had been in the US market for three years by that point. And it was starting to make the pivot from being a very, a platform that assumed that its user had very clear ideas of what music they want to listen to, to trying to reach a more mass audience. And so

Part of the difficulty we're having at that point was, you know, a lot of listeners don't know necessarily what they want to listen to, right? This is the genius of radio is you just turn it on and it's already curated for you. And so this is the period when, you know, not just Spotify, but a lot of platforms were starting to design, you know, podcasts.

different kinds of playlists that you could just click on and listen to and not really have to exercise any choice there or recommendation algorithms of various sorts that would say, oh,

you like this kind of music, we can recommend you some more things along the side. And then from there, right, it just, I started to get fascinated by the whole streaming ecosystem and what its sort of implications are for music culture. Yeah, absolutely. I mean, kind of building off that point, you know, you're saying your curiosity being piqued about who the audience for music information retrieval research is, but I suppose quite a central question is,

and knotty question that animates much of the book is who is the audience in general, right? Because the point that you make quite consistently is that the straightforward answer, perhaps, that the listener is the audience for the music that streaming platforms supply is not the only question. So I wondered if you could speak to that a little bit about some of your findings there. Yeah, well, I mean, you know, so I think...

This gets to, if I'm understanding your question correctly, I think one of the things that's important to bear in mind about platforms, and this is an observation that scholars of digital platforms have made,

made for many, many years now is that the design of platforms is one where there are many different kinds of audiences or user groups, right? The sort of technical term in the literature on the economics platforms is that these are multi-sided marketplaces. And the idea is that you leverage

one user group off of another in order to gain what, to use Parlin's network effects, sort of virtuous cycle where

The classic example is a sort of napkin that the founders of Uber drew their business model on, where the idea is that the more drivers you attract into the network, the more passengers you then are likely to attract, and so on and so forth. And before you know it, you have a very viable business model that's scaling rapidly because the cost of adding one more driver, one more user to the system is negligible.

So when we think about streaming and music streaming more specifically, the same kind of logic applies in broad brushstrokes, even though obviously there are clear differences between, say, a music app and a rideshare app and other kinds of platform-based apps. So in music streaming, you know, you've got the listeners for sure, right?

And they're one user group, but not necessarily, certainly not the only one and not necessarily always the most important one in terms of the economics of platforms. You have also artists, which interestingly are defined as, you know, a different kind of user group, which is,

If you're working with a more conventional kind of image of the way music industry works, we would think of it more as a pipeline, I would think, where you have producers at one end, consumers at the other end. But the platform is basically treating these as not necessarily equivalent, but they're treating them both as distinct user groups that they're mediating the interactions between. Yeah.

That said, artists are usually... There's one layer of mediation separating the artist from the platform, and that's either their record label, if they're assigned to a record label, or some sort of distribution platform like DistroKid or CD Baby or something like that. Spotify, you can't directly upload your music to the platform if you're a musician. You have to work through some sort of intermediary. But other audiences include...

which, you know, vary from platform to platform, but many platforms have a sort of ad supported or freemium tier. And so, you know, obviously they're feeding content in the form of advertisements to platforms, but they're also consuming content after a fashion. It's just the content they consume is, you know,

you know, data on listeners on their demographic profiles, on their psychographic features, things like that. Um, and,

And then, you know, the fourth user group that's very important for certain platforms, but not all of them are investors. So this is certainly the case with Spotify. And this is, you know, one of the most important audiences for a lot of Spotify's communications is their investors. You know, they went public in 2018. Prior to that, they were

funded by various private investors. Now they're listed on the New York Stock Exchange. And so if you go to their website, they have different portals for all these different user groups. And so there is a sort of Spotify for investors page where you can watch videos, download earnings reports, read the letters that the CEO sends out, things like that.

So, yeah, so defining who the user is on these platforms is a tricky thing because it's a multi-variegated entity. Yeah, and I think you do a really good job of never taking for granted any of those questions about intended audience, even, you know, prices of recommendation or curation that seem straightforwardly

a part of a platform's key offer to its listeners you often kind of get into why that might not was not it might not be quite so simple right and i think uh a key dynamic there i suppose you and you touched on it you alluded on it in terms of the question of advertisers is the fact that

As much as there's the obvious question of serving ads in amongst the audio on a freemium tier, it's slightly more complicated than that. And that's what much of the data capture going on is for, is this kind of profiling process. So I wondered if you could explain that a little bit more, particularly how it relates to platforms' kind of process of assetization, the process of data capture and profiling.

if that makes sense. Yeah, yeah, yeah. I mean, so, you know, I think that one of the things that comes up a lot is, so, okay, so I'm going to talk a bit about Spotify specifically, though. I mean, a lot of what I have to say will apply to other platforms. So, you know, one big distinction that's worth making from the get-go is that

You know, we can make a broad analytical distinction between standalone platforms like Spotify, Deezer, there are others out there, Tidal. Yeah, Tidal was acquired by Stripe, so I don't know, maybe their status is a little bit more ambiguous. But...

So we have standalone platforms on one side, and then on the other side, we have those that are streaming services that are sort of bundled in with larger tech companies. So Apple has their streaming platform, Amazon. Google has its sort of premium YouTube music platform. So a lot of what I'll say right now really addresses more of the standalone platforms, and a lot of what I'll say really focuses on Spotify. But like I was saying, this is...

stuff that also will apply to a lot of other platforms. But, you know, the thing about Spotify that's interesting is that it's only been in the past year that they've actually started to generate a profit. And this is something that many people pointed out that, you know, on the one hand, you know, from a kind of common sense, you know, everyday common sense point of view, you know, it's sort of

you would think that the point of a business is that you earn money, but, um, you know, Spotify had been around for about 16 years and, you know, hadn't actually earned a profit, um, except for like the odd quarter here and there. But, but for the most part, you know, if you look at their revenue, it was going up year after year. Um, they're bringing in a lot of money. Um, you know, one of the problems of their particular business model is that they are so dependent upon, um,

Right.

And so, you know, that's a fairly punishing economic or business model to be working with. They've made some changes the past year, which has allowed them to generate a modest but, you know, significant profit. But, you know, then the question is, well, how is this company staying afloat? And

From this perspective, what they're doing is not at all unusual. This is sort of the Silicon Valley sort of tech startup model, which is initially get a lot of private investment from patient investors who are assuming that you are going to scale so rapidly, you, the sort of tech startup, because of the sort of virtues of network effects.

Um, and therefore corner a market dominated. Um, and then at that point, you know, you will be in a sort of, um, you know, you'll have the whip hand and be able to set prices and you'll be able to, um, become profitable in the way that like say Google and Facebook and Amazon.

And so, you know, you get this like interesting disconnect between Spotify's market valuation, you know, which is quite, quite high in their, you know, earnings, which are, you know, negative or have been until very, very recently. And so what's to explain that disconnect? Well, I mean, part of it is like the assumption that at some point Spotify is going to dominate the streaming market and they're going to generate profit.

But I think part of it also is that, you know, Spotify is sitting on the stockpile of data, right? That's, if anything, their most valuable asset. You know, if we look at what Spotify actually owns, they own very little.

You know, they don't own the music they stream, right? That's owned by rights owners. You know, they actually don't own much of the infrastructure through which they stream music because they like a lot of, you know, I'll use this adjective guarded, but we smaller tech company and Spotify, it's a big company, but relative to...

It's a pretty modest-sized company, but like many companies of its size and smaller, it leases a lot of its employees.

data storage, its computational processing facilities, all of that from Google in the case of Spotify. So what do they actually own? Well, they own some patents, which isn't to be sniffed at. That's valuable intellectual property. And they own all the data that they've sort of collected on users over the past 15 years. And that

Given the sort of importance that data has come to acquire in the digital economy, as a number of scholars point out, it's almost tantamount to a form of capital, right? This is the argument that J. Thadowski has notably made.

And so, yeah, so this is one reason why Spotify is as valuable as it is as a company. And it also explains why they are so aggressive, like a lot of tech companies in the collection of data, right? Because this is a valuable commodity that they can monetize in various ways.

So what's interesting to me in studying this is like the way they position music relative to these data collection practices. Right. Because a lot of, you know, we were talking before the podcast began about Google Chrome. Right. And every time I turn on Google Chrome, it asks me, you know, can Google access music?

you know, all these files on my computer, I have to do it if we say no, no, you can't. You know, so this is pretty common, right? You know, like tech companies are famous for being very, you know, very, very greedy when it comes to data. What's interesting is that, you know, the way in which Spotify figures music into that kind of practice, into the narratives it builds around its data collection, right? And so, yeah,

This really becomes apparent if you look at the sort of communications it directs towards prospective advertisers when they're trying to essentially market themselves to marketers and say, like, why should you work with Spotify? What's the advantage? What can we offer you that, say, Facebook can't or Instagram can't or TikTok can't?

And, you know, what's interesting there is that, you know, they structure a lot of their appeal around these arguments about what music is good for and the ways in which it reveals certain

intimate details of people's personalities, their moods, their activities, all these things. And they're really drawing on like discourses that I would argue go back decades, even centuries, right?

that relate music to, you know, the interiority of subjects and sort of say that, like, music is a sort of reflection of our, like, deepest being in some profound way. But also, you know, the notion that with the development of music

various kinds of portable music devices over the past hundred years, the latest iteration of which is the smartphone. You know, music is this incredibly portable thing that we take with us and

that accompanies us, that's almost like a companion for us. And so basically the argument is that, you know, music can be with people and animate their lives and shape their moods in places and spaces and occasions where other media can't. So, yeah.

One of the examples I cite in the book is the shower playlist. You can't really watch a video or scroll through Instagram in the shower, but what you can do is have either your phone or some kind of smart speaker set up in your bathroom and listen to music while you shower. This is a number of years ago, but I think it was in 2018, one of the marketing projects

executives at Spotify made a point of saying that at the time there was something like 40,000 user-made shower playlists on Spotify. And so, you know, that sort of, from their perspective, is very, very attractive or should be for users.

advertisers and the various kinds of data aggregators and data brokers that sort of mediate the market for digital advertising, because basically this gives you real time information about what this person is doing at that moment. You know, if they're listening to a showering playlist, then presumably that person's showering. And that can be valuable because then you can target and add to them at that precise moment for shampoo or whatever.

And with Monday.com, work is just easier.

Monday.com for whatever you run. Go to Monday.com to learn more. This is the sound of your ride home with dad after he caught you vaping. Awkward, isn't it? Most vapes contain seriously addictive levels of nicotine and disappointment. Know the real cost of vapes. Brought to you by the FDA.

Yeah, absolutely. And I guess going back to how you kind of got into this work, I wondered whether you might talk about what this renewed, perhaps renewed or novel emphasis on context and on the unique capacity of music to kind of follow users around, be simultaneously everywhere and also deeply internal, you know, offering unprecedented capacity for music

at least the story goes, surveillance of a user's inner emotions, what that kind of emphasis might be doing on earlier forms of musical categorization like genre. You know, how does genre live on in the streaming context and how has it maybe changed? Yeah, I mean, that's... Yeah, and that... I mean, thank you. I mean, that is a fascinating question. I think, you know, obviously...

So with a lot of this, right, I mean, it's really hard to say where we're going to land. I mean, this is, you know, the streaming ecosystem, the streaming economy is a moving target. And I'm loathe to make predictions.

hard and fast pronouncements about this is what genre is now because who knows how durable these changes are. But I think one thing that you can do is go to any platform and look at their own categorizations that they promulgate and they're revealing. Yes, you will have on a lot of these services playlists organized by what are recognizably

familiar genres like R&B, hip-hop, classical, jazz, country. You know, it'll vary from country to country depending on where you are.

I'm in Texas, so when I go on to these services, a lot of times there will be a playlist category for regional Mexican music, which reflects the types of patterns of listening that are typical to my audience.

location. You know, if you go to Spotify, Pakistan, presumably you're going to have more Pakistan music there. So yeah, so you have those traditional categories, but alongside them, and this is what's interesting, is that you see a lot of new categories. I mean, sometimes these are

things that are related to specific properties that Spotify's invested in. So one of the few things that Spotify does own is what?

certain podcasts where they have exclusive contracts with certain podcasters. So for a few years there, they had an exclusive contract with Joe Rogan, the most popular podcaster in the United States. So, you know, there's like a tab for Joe Rogan. But, you know, alongside, you know, the traditional music genres and those like kind of weird, highly specific categories, you have things like mood,

or chill or sleep or dinner or what else um exercise various workout playlists um and so you can see how these categories which you know it's not as if like music hasn't been around for a long time or it's not as if there haven't been you know exercise cds that were um

you know, sold in record stores back when the sale of physical recordings was at the center of the music industry. But, you know, from my experience, right back in the, say, 1990s, if you want to buy a CD of like workout music, there's like a bin in the back of the store where it's kind of cluttered.

you know, jumble of, um, you know, workout recordings could be found. It was not, uh, um, listed at the same level as say the, the rock or the, the hip hop section of a record store. Um, but you know, basically one thing we see on a lot of these platforms is that, um, categorization around mood activity, um, and, um,

context is becoming more and more prominent. And as the years have gone on, these have become increasingly hyper-focused. And partly this is a sort of way in which a lot of streaming platforms are in a kind of dialogue slash competition with, say, TikTok and other kinds of platforms where you have these

micro trends that will come and go. And basically a Spotify will try to latch onto that. So, you know, one of the sort of playlists that, you know, was on Spotify a few years ago, I assume it's still on there as pumpkin spice, which is an autumnal cold weather playlist. I mean, and what goes into that playlist is,

Yeah, I mean, it's a good question, right? You know, how are these, you know, various kinds of, yeah, attitudinal mood-based characteristics getting mapped onto music? I mean, a lot of this seems, from my perspective, kind of arbitrary. I mean, there is a logic there for sure. But, you know, one thing that I think that the streaming economy really incentivizes is,

is, you know, the organization of music in these terms, because those are the terms that are economically valuable to streaming platforms, since one of the things they're trying to do is use music as a technology of surveillance in order to track people's moods. So it makes sense that they want to categorize music in that way and get people to think about music.

in those terms. Yeah, and it's really interesting and you make this point very effectively that while there's this quite significant shift in terms of emphasis on categorization and the presentation of music, the older ways in which genre and other forms of categorization associated or for which genre was often a proxy

proliferate in music particularly regarding kind of demographic identifiers and special groups also still persists and maybe proliferate in the kind of the context of algorithm recommendation and streaming right so um you know race records don't disappear for example uh yeah and there's i mean there's a great uh music scholar tom johnson whose dissertation did some like wonderful empirical work showing the ways in which um you know the sort of seemingly innocent

you know, tags that are assigned to different musics are very much determined by the sort of, you know, putative racialization of various artists. So, you know, white indie artists tend to accrue more tags. And if we think about this in terms of theories of like cultural omnivory and that as a certain kind of status marker that suggests that, you know, there is like a greater status being assigned to

Those artists then say comparably, you know, eclectic artists, the example he cites is Rihanna, who tends to not accrue as many genre tags and therefore sort of symbolically devalued her music is.

I think, you know, one of the interesting things as well is, you know, what happens to this data that, you know, say a Spotify, but other, you know, streaming platforms will collect and then through their commercial platforms,

interactions with various kinds of advertisers and data aggregators and data brokers, how does that then sort of work its way into other kinds of classifications outside the streaming ecosystem?

And, you know, I think this is a critical question. It's a hard one to address. And it's one that, you know, I try in the book to, you know, couch in very careful terms because ultimately we can't very easily track where our data goes anywhere. There's been some very helpful research.

Empirical work that's been done that tries to show all the data sharing agreements that different digital platforms enter into with one another. You know, there are certain economic incentives that lead platforms to enter into that kind of data sharing agreement. One of which is that as of now, data has this kind of nebulous status within standard accounting procedures. So, yeah.

The sort of system of national accounts, which is sort of the international agreement for setting accounting standards globally, is actually undergoing a kind of revision of its standards. And presumably when the new standards are released next year in 2025, data will be listed as a new asset class.

and there will be guidance given on how data assets should be valued. But, you know, that hasn't been the case until now. And so as a number of sort of business scholars have pointed out, you know, this makes data a very attractive thing for companies to share because it sort of doesn't appear on the books, you know, and so therefore, among other things, it's not really taxable, right? You can sort of trade data, but not, you know, incur any kind of like liabilities from that.

And so, you know, these platforms tend to be leaky, right? So in order to offer targeted advertising, usually, you know, platforms have to work with data aggregators of various sorts. Those data aggregators, you know, one of the things they're trying to do is essentially assemble these profiles of individual consumers that will, you know, basically follow you around the internet, right?

no matter where you go, so that you can be sort of targeted more effectively as an individual. And so that means that there are these consumer databases, and this is particularly the case in the United States, where it's

fairly unregulated, the data market. Obviously, the EU has slightly more stringent regulations governing the use of personal data than the United States. So, you know, the degree to which, you know, people living in the EU are

uh, have to deal with the same sorts of, uh, consumer classifications that say citizens of the U S have to, to sort of deal with, uh, it's an open question. Um, but you know, in the U S at least there's, um, yeah, these consumer databases where like essentially they have like countless scores for all kinds of behavioral attributes, um,

In order to figure out, you know, not just, you know, what sorts of goods and services to market you, but also these are of great use for insurance companies, various stripes, you know, and trying to determine risk assessments on individuals. Banks may not use some of this data, but, you know, they're all these now, you

you know, FinTech companies that offer sort of alternative financial instruments to, you

various people who either want some alternative to, say, a credit card or a bank account, or people who, for various reasons, because they don't have a credit record, can't get access to credit. And so they have these alternative forms of data that they use to judge creditworthiness. And I don't want to overestimate the value that

You know, music data in particular has because, you know, it's just going to be one kind of data among all the other kinds of data that we sort of give off in the course of our everyday lives moving through digital spaces.

But, you know, at the same time, it is, you know, almost certainly one of the forms of data that sort of figures into these profiles of people that are put out. And, you know, one of the interesting kind of places where you see this, you know, hints of how this might have bearing on people's lives outside the streaming platform or things like that.

Automotive insurance, which has a long and not particularly happy history of offering different people different rates based on.

zip codes, which is usually in the United States a function of, you know, proxy for class and race. And, you know, there have been regulations have been put in place to try to sort of limit insurance companies' ability to do that. But, you know, you can sort of see how music can function as a proxy for various kinds of

uh, racialized categories and therefore various kinds of racialized risk assessments. Uh, so one thing I cite book is, um, you know, uh, uh,

put out by a brochure put out by the insurance company Allianz about 10 years ago, where it basically tries to say which kinds of music are associated with riskier or safer driving. Perhaps unsurprisingly, like classical music and sort of middle of the road, the easy listening is associated with safer driving. And then what's associated with riskier driving is hip hop associated

essentially. And, you know, they provide an infographic to illustrate this and it sort of makes the sort of racial coding of these categories, risky versus safe, abundantly clear. But, you know, I think that this is just one example of how like data gets recycled and put to all kinds of

potentially very problematic uses. And so that kind of argument that the music we consume is especially revealing of who we are, both demographically, but also psychically in a certain way, you know, is perhaps true to a certain extent, but it also is clearly being instrumentalized in

in profoundly problematic ways by not just streaming companies, but other actors within this broader data economy.

Yeah, I mean, I found this section so rewarding because I confess, you know, I, along with most people, am concerned about various forms of digital surveillance, but music has always seemed to me like the kind of relatively, you know, trivial concern compared to most. But then, I mean, yeah, that image in particular from the alliance insurance thing that's mapping genres onto dangerous and safe, I mean...

It's not even a dog whistle, right? It's so explicit and it's so recent. It's 2012, right? It's kind of amazing example of the role that supposedly, I guess, the ideology of data being implicitly neutral can be put such immensely dubious ends. And I suppose that a point that occurred to me, as you were saying just then is,

Because you mentioned in passing about the kind of sale of third party data to law enforcement and how that's quite prevalent in the US and in the UK. And I know in the US as well, there have been a number of instances of prosecutions on the basis of listening to certain music, invariably hip hop or in the UK kind of drill, right?

on exactly those terms right so it kind of your your work here really reveals the stakes are are kind of that high or can be that kind of high even with music you know yeah there's i mean just to cite one thing that doesn't occur in the book but it's something that um i'm still doing work on um music data collection and some of the uses it's put to but um

Continue with the insurance example. So State Farm Insurance, which is a large insurer in the United States, they do home insurance, they do automotive insurance. I mean, I'm a State Farm customer. I guess it's okay for me to say that on the air. This isn't an endorsement. It's just...

Factual representation. But, you know, one of the things that they've been doing along with a lot of other auto insurance companies is really pushing their customers hard to install a telematics device in their car that will track your driving in real time.

And the argument they make is like, listen, if you're a safe driver, this will tell us how safe a driver you are and you may see your rates go down as a result. You know, they don't say the obvious inverse of that, which is they can also raise your rates. But the thing that's interesting is like the patents for that State Farm has filed with the U.S. Patent Office for such a telematics device is

are really make for interesting reading. Now, you know, whenever you talk about patents, you have to sort of use a bunch of caveats because not every patent filing that is, you know, published, you know, actually translates into a marketable and marketed technology. Sometimes these are like purely defensive or aspirational patents.

But nonetheless, State Farm does have a telematics device that seems to match pretty closely in a lot of its particulars, this device that's described in this patent that they put a filing in for. And they talk about all the data they collect and write obvious things like,

What speed are you going? How hard do you brake? Are you swerving? Things like that. The obvious kind of driving day you would expect us to collect. But then among other things, it's like, what is the music you're listening to?

and they have some kind of machine listening algorithm built into this device that's going to, first of all, maybe using audio fingerprinting or some other technology, identify the kind of music you're listening to. But it's also going to identify the tempo, the loudness. And in the patent filing, it's sort of saying, well, let's say this person has...

uh, habit of listening to, I think the example they use in the patent doing this from memory. So I may be getting some of the details wrong. They say it's like this person has a habit of listening to acid rock, which has this average tempo, this average loudness, um, from this, we can deduce or infer that this person, because people listen to this genre of music tend to be like

I think is the term they use that we can create a kind of personality profile, this driver, which is going to, again, feed into our assessment of this. Is this person likely to be more prone to getting into accidents or being an unsafe driver in some way, shape or form? And then that can then feed into that.

You know, rates that will just almost in real time as well, right? So your rates could go up and down, like from a month to month basis, and like music is feeding into that determination in a very kind of direct way, according to this patent filing.

Gift the remarkable with Marc Jacobs fragrances this holiday season. From the iconic Daisy and Perfect to the all-new Daisy Wild, Marc Jacobs perfume gift sets include everything she needs to feel special. From her favorite fragrance plus the matching travel spray, holiday gifts don't get much more perfect than this. So if you're looking for a gift inspiration these holidays, gift the remarkable with Marc Jacobs.

Just beautiful.

Whatever you love, find it on eBay. eBay. Things people love. Which, and presumably the kind of underlying models by which these data points are ascribed to a set of sonic characteristics and then...

made, what's the word, made to correspond with a demographic profile, you know, can't be too far from the same sorts of data capture techniques and modeling used by streaming platforms, presumably, you know,

Yeah, I mean, it's interesting. I mean, it is. I mean, like one of the challenges in reading Platin says that they give, you know, they describe the technology without obviously revealing any trade secrets. So who knows exactly what technology they're using, if they're contracting with some third party company.

music data analytics firm to sort of analyze this music that you're listening to using machine listening algorithms. My guess is that State Farm is not building from scratch their own machine listening tools. But yeah, I mean, at least according to the patent filing, they're going to

the music by genre, make some basic kinds of observations about, you know, music characteristics like tempo, loudness, I think even like key maybe. And then from that, essentially, you know, try to track what you're doing in real time, you know, and see if, deduce from that whether or not they think you're likely to be driving fast

And then, you know, just as problematically trying to build a sort of longer range personality profile. Right. So, like I said, you know, in this patent filing, they just ascribe a certain kind of.

personality type to acid rock listeners you tend to be mildly depressive or something like that if you listen to charm but you know like yeah maybe uh i i i personally would be depressed if i had to listen to nothing but acid rock but you know um you know that's just me uh i'm not making decisions about you know how much or how little you have to pay on car insurance so

I guess that speaks to an underlying issue and an underlying kind of problematic issue that runs throughout the book, which is, as we've kind of already touched on, the idea that a stream that comes through the platform is commensurate, A, with any listening at all, and B, with a then stable or at least detectable personality profile, which comes to the kind of problem of attention, which is what you talk about earlier.

most specifically in your chapter on kind of the notion of fake streams or and so I wonder if you could talk about that a little bit what what kind of problems and solutions are raised by this um

between the stream and the listen or the stream and the underlying attention? Yeah, I mean, it's quite interesting because I mean, I think that, you know, the entire streaming economy is predicated on the simple assumption that like one stream that's registered, you know, when you press play on a track, it corresponds to some act of listening and

That, you know, granted is going to be defined in very broad terms and loose terms by platforms. But still that, you know, there is some kind of correspondence between, you know, that action of your computer registering a play, that information being sent back to the platform servers and that there's corresponding to that some, you know, active listening, other passive or active engaged or disengaged,

um, on the part of the user. Um, now clearly there's a gap there, right? Um, you know, there's no sure way to know that, you know, a play equals a listen. Um, and it may be that you put on a playlist and then like went out of the room for a while, much like you might leave the lights on, um, in a room and, and forget about it. And then the

And we realize when we come back in half hour, 45 minutes later, you know, there's no way that the platform knows whether or not you're physically present. And even if you are physically present, they have no way of knowing like in what way you're attending to the music or not attending to the music. But, you know, there has to be that fundamental assumption that, you know, there is some level of engagement because, you

This is what advertisers are paying for. This is what

Rights holders are, you know, I mean, they're kind of paying the platform. I mean, they get revenue from the platform, but, you know, the platform gets its cut of that revenue, which is in part, you know, payment that rights holders are making to the platform for, you know, performing a number of different services on their behalf, one of which is to distribute their music to audiences and, you know, through that generate engagement with this music.

So that gap between the sort of representation and whatever experiences on the other side of the interface

There's one that, you know, I think it's most, uh, vividly expressed in the forms of streaming fraud that I talked about in, in chapter four. Um, you know, this is something that, you know, when I first started doing work on this book, you know, very few people were talking about this has since become like a major issue within the music industry. Um,

You know, there are now specialized startups. They're all about detecting, you know, fraudulent streams, things like that. You know, it's one of these things where, you know, I think that the way that it sometimes gets talked about in the music business press is troubling because, you know, I think that the working assumptions that there are like these

good, honest, hardworking musicians who are being kind of taken advantage of by, you know, scammers. And, you know, there are scammers. I mean, there's a famous story of like, I think like some Swedish mobsters gaming Spotify and making, using it to launder money or something like that. But, you know, in a lot of cases, these are musicians who are trying to

figure out a way of getting their music heard in a, uh, within a system that's very, um, very much designed in order to advantage major label artists, um, who have all kinds of structural advantages in terms of, you know, the playlisting of their music, um, having certain kinds of, um, increased promotional access, um, uh,

you know, having, you know, just a larger claim on the revenues that streaming platforms generate through, you know, the, the various kinds of licensing deals that they've struck. Um,

And so, you know, I think that, you know, in some cases, you know, these efforts are just done by musicians in order to try to break through in some way. So, you know, but regardless, all of them hinge upon exploiting that gap, right, between, you know, the digital representation and the active listening. And, you know, some of them do so by, you

you know basically engaging in various kinds of click fraud so this could be using bots to artificially run up um you know the stream counts for a given track it can be a matter of paying click workers um you know uh in various parts of the world though

As with other kinds of Qlik work, a lot of this work is concentrated in the global south. When I was doing research on this chapter, things have changed substantially since the late 2010s, but there are a lot of micro-work platforms where

People involved in micro labor, primarily centered in places like Bangladesh or the Philippines, could basically advertise to potential customers, pay $5 and I'll get you 10,000 streams, pay $10, I'll get you 20,000 streams or something like that.

And you could contract with this person if you were a musician. And there are a lot of posts on the same sites from people

not on the part of sellers of their labor, but from, you know, customers from buyers of, you know, micro work, you know, musicians who are saying like, I want to break into the charts. I need some way to, to, to boost my, my plays. I'm willing to pay this much. You know, some of that has been sort of clamped down because that was obvious, you know, obvious sites that, that, you know, could be, you know,

um regulated out of existence basically but you know basically people have moved on from there um there are different ways that you can artificially inflate streaming numbers on the flip side the other kind of possibility is you know to use various kinds of um

shortcomings or deficiencies with streaming platform interfaces to try to, you know, either dupe users into clicking on a track which you might not, you know, otherwise click on. You know, I use a lot of examples and, you know,

the book of knockoff covers, which, you know, kind of pass themselves off as the original track. And, you know, in some cases show remarkable talent and artistry and their ability to emulate the sound of the singer's voice, the production of a backing track. These are sometimes very impressive moments.

impressive productions and they're often quite ingenious um you know i mean kudos to some of these people uh for thinking through how to get people to listen to some tracks like one of the examples i use is um

Right.

Right. But what these artists realize is that, well, what we can do is just give you the full title of or the full line from the chorus. I'm only human after all, because that's probably the person who's in the checkout line at a supermarket or in a coffee house. And here's this in the background. They hear that chorus. They think that must be the title of the song. They go to Spotify. They look up that title. Right.

what turns up isn't the song human but rather all these other alternative knockoff cover versions and you know they do reasonably well i mean i think when i was doing the research some of them had accrued like five million streams which is you know more than a lot of you know um tracks by you know artists who you know otherwise have no profile um

You know, things, I think one of the things that's like interesting is how these kinds of ploys develop over time. So, you know, the thing that a couple of years ago was big was tracks that just said they were going to have a featured artist who is somebody famous. So, you know, some hip hop track by somebody you've never heard of, but it's featuring a

you know, Kendrick Lamar. And just the fact that they could put that there meant that, you know, you're going to get a certain amount of traffic to your track and hopefully get people listening to it. I mean, there are also people who are uploading playlists of movie soundtracks and like surreptitiously inserting their own music in between tracks and, you know, getting people to listen to their music that way. You know, so part of this is, you know, the sort of

of people who are trying to just get by and have to devise these kinds of tactics in order to direct attention to them, like I said, within a system that's very much geared towards funneling direction towards the biggest superstar artists. I think what's so interesting about this and that you show really well is that...

Well, there's a bunch of things. One on the kind of subject of the rag and bone man that occurred to me is that it's a kind of recursive optimization where, as you detail earlier in the book, artists are being kind of structurally incentivized to be very chorus heavy, very hook heavy, songs are shorter, they introduce a hook earlier.

And as a result, your average listener is much less likely to know many details about the song, bar this very hooky line. So then the counterfeiters are much better placed to optimize their fakes to be picked up. Do you see what I mean? By reading their song after that thing. So there's that. And then the other fascinating thing, I suppose, is that

arbitrating what is illegitimate or legitimate on the part of the platforms so often uses a kind of moral or ethical register and as you kind of suggest not least because the boundaries between legitimacy and illegitimacy are very fuzzy a in terms of the way the platforms themselves are kind of organized and orient their own practices but b legitimacy is socially and culturally contingent so i wondered if you could speak to both of those things

Yeah, well, certainly, yeah. I mean, to a certain extent, I just want to say yes. You know, like legitimacy is like culturally specific and it's a mutable variable term. I think that, you know, this ties back to like the sort of contingency and mutability of attention and like what counts as attentive and non-attentive. And

Yeah, I mean, platforms have a strong incentive to try to blur those boundaries up to the maximal point that they can sort of get away with, right? Because, you know, it's in their interest that, you know, they encompass the widest variety of, you know, options.

listening behaviors right ranging from like the the totally disengaged person who just has you know kind of sonic wallpaper on might be even something like just white noise to sort of drown out ambient sounds they can focus to somebody who's you know like i don't know like music scholars like us you know who you know we are professional deformationists to like care deeply about music and listen attentively to it and and think critically about it all the time um

But yes, I think that part of the trickiness then for...

platforms is that, you know, what counts as like legitimate listening and legitimate music also varies according to like listening communities. Right. So one place where you see this crop up a lot is in, you know, community forums where people either advocate for or against like a greater share of karaoke tracks on, um,

You know, platforms, you know, for certain people, this is a kind of an abomination. It's not really music. It's something else and it shouldn't be on the platform at all. And, you know, these are people who, for whatever reason, they got some karaoke version of the song they want to listen to instead of the actual music.

and they get very upset by this. On the other hand, obviously, yeah, people like doing karaoke. And for them, this is something that there should be more of, not less of on these platforms. And so this is something that, you know, by sort of becoming the sort of de facto thing

hosts of the world's music and essentially tried to aggregate, um, not just music at this mass scale, but audiences at the mess, this mass scale, um, you know, platforms find themselves in this, uh, difficult position, uh, of their own making where they're trying to, um, balance, you know, these competing definitions of, uh, music,

legitimate music making and legitimate listening. I think one of the things that's become increasingly clear in the past year or so, and this is something that didn't really figure into my book just because it was only a

really sort of coming into being right as I was doing the page proof. So it was too late for me to really address it in any kind of like deep way, but there has been a shift and the way revenue is shared on certain streaming platforms. And this is coming at the behest of the major labels in particular, universal music,

which in, I think it was January of 2023, the CEO of Universal, Lucien Grange, you know, published a letter and basically said it's time for the streaming economy to undergo an overhaul. And the thing that he was denouncing, among other things, were various kinds of fraudulent streams, but also the ways in which, you know, what he and UMG by extension were categorizing as sort of basically fraudulent

non-music or not legitimate kinds of music making or, uh, you know, uh, devaluing, you know, the, the, the sort of legitimate music by professional artists on the platforms. Um, and so essentially they propose this new model for sharing revenue, which, you know, it is basic kind of mechanisms is similar to the old model that, that, um,

Most platforms had been using since the outset. I'll try to keep this brief because it's one of those things. You go into too much detail, people's eyes will start to glaze over. But basically, the way the system worked in the past was that a company like Spotify or Deezer would collect all their data

you know, your users, you know, revenue that they generate from users put into different pools depending on like which service you're subscribed to, premium versus premium.

And then essentially, you know, apportion that to rights holders on the basis of like what percentage of, you know, the streams did your artists, uh, generate over the course of like a reporting period. Right. So if like Taylor Swift is responsible for like 20% of the streams on Spotify and a particular month, which, you know, who knows? I mean, it's Taylor Swift. It could be possible. Um,

Then her record label gets 20% of the overall revenue that they pay out minus whatever Spotify's take is. And then according to what are the terms of the contract that Taylor Swift has with the record label, she gets a certain percentage of it.

So the new model basically keeps that system intact. But what they do instead is say that in order to become eligible in the first place for receiving any royalties whatsoever, you have to reach a certain threshold, which is a thousand streams. And they have to come from at least 500 distinct users. Right. So it can't just be like a single user.

you know, I mean, one of the sort of classic scams is to sort of set up a playlist and set up like a handful of accounts that just auto play, you know, a playlist over and over and over again. And,

So it's trying to, you know, at least nominally get rid of that kind of streaming fraud and sort of say, okay, the cutoff between like legitimate music is this 1000 stream threshold. Now, this is obviously arbitrary, right? Who's to say that 999 streams is, you know, qualitatively distinct and illegitimate, you know, whereas like you get that one more stream now and you're a real artist, right?

And, you know, the problem is that obviously there are lots of genres of music that don't generate a lot of streams for various reasons. And, you know, this could be certain kinds of like avant-garde music, noise music, classical music, jazz, right? These are not things that are like big on streaming platforms. Are these people not real musicians? Yeah.

because they didn't hit that artificial threshold that the major labels are imposing. No, obviously not. But, you know, again, this shows us where the power in the streaming economy lies. It really lies with the major labels who own the assets that, you know, a platform like Spotify owns.

needs in order to function in order to continue to um you know attract users um if universal were to pull its catalog from spotify that would be maybe not a death blow since you know they did that to tiktok in the spring and tiktok managed to you know scrape by until universal finally you know agreed to put their music back up onto tiktok but still

Spotify over the long term isn't sustainable if the major labels pull their catalog. So that means that the major labels have a lot of ability to sort of set the terms of how these companies run. And the fact that essentially Deezer very quickly said, yes, we like this new system. We are going to adopt it. Spotify followed suit.

shows you that essentially Universal in particular, but the major labels in general, are able to really sort of say, okay, this is how we want revenue to be shared. And this is by extension how we're going to define what counts as real music, what counts as a real musician, and this is what's going to be shunted off. Now, one last thing before we leave the subject behind is

It's important to bear in mind what happens to all that revenue generated by the artist who just makes it underneath that threshold, or anybody underneath that threshold for that matter. Well, basically it gets lumped in with the revenue pool that then gets allocated to those artists above 1,000 streams. So essentially what this means is that this is a means of

Sure, maybe rooting out some streaming fraud. Okay. But, you know, probably the more significant impact of this new system is to redistribute income upwards in a pretty profound way, right? Because essentially all those artists, and there are millions and millions and millions of them who don't hit that threshold, they're going to be demonetized. All their money is now going to be reallocated to, you know, those artists who meet that criteria.

Streaming now on Peacock, Paris Hilton and Nicole Richie are back. That's hot. Loves it. For a show-stopping reunion that will prove putting on an opera is anything but simple. We're really good at this. One thing's for sure, they won't be upstaged. Good to have you back. Come on, we've got a show to do. Paris and Nicole, the encore. A three-part reunion special. Streaming now, only on Peacock.

This year, Santa's bringing the power of Energizer into his workshop. Whoa, the Energizer bunny's got so much power. Wait, he's powered up all the toys. I think that means we're done for the year. I love this bunny. He's the hardest working helper the North Pole has ever seen. And he wants all your gifts to have the power of the number one longest lasting AA battery. So this holiday season, stock up on Santa's and the elves' favorite battery, Energizer Ultimate Lithium.

And I suppose that quite neatly takes us onto the notion of expropriation, which is a, like quite a key theme, particularly in your, in your final chapter. And I should say, you know, that the, the use of expropriation, it speaks to Eric's kind of use of very like refreshing use of a Marxist approach throughout the music industries, which we haven't touched on the nuts and bolts of so much, but in this final chapter,

you particularly use Marxist feminist approaches to social reproduction and

and the question of expropriation is quite a live one here and you do this quite real really interesting move that compares the kind of the broader use of cheap things you know cheap labor cheap social reproduction and cheap nature to the approach that the streaming economy takes to musicians and music workers so I wonder if you could explain that as it's quite a big topic but it's such a fascinating chapter in the book

Yeah, thanks. I mean, one thing I would say is, you know,

when we were talking before the podcast began about like some of the difficulties in writing a book like this. And, um, you know, one of the challenges is that, you know, the streaming economy is changing all the time, as we were just talking about before, you know, the revenue share model, uh, seems to be evolving, um, you know, and other changes will no doubt take place over the next few years to come. So trying to write a sort of definitive statement was always never really in the cards. Um,

But, you know, I think that one of the things that became clear to me in writing the book is that streaming offers a kind of useful moment to reassess, you know, how we think about music's relationship to capitalism. And one of the things that...

You know, I think streaming throws into stark relief is that music occupies a very ambivalent relationship with respect to what we might refer to as the formal economy. Now, sure, there's things like the record industry. There are things like copyright law, which is, you know, basically a means by which, you

you know, the sort of public good characteristics of music, the fact that it's very hard to exclude people from, you know, enjoying music, it's very hard to, you know, it's non-rivalrous, you know, my enjoyment of music doesn't, you know, diminish other people's enjoyment of music. You know, copyright's a way of, you know, basically converting what otherwise might be a public good into a private good, which is more amenable to commodification. So there are various kinds of

technical fixes that you can come up with various kinds of legal fixes, which can like make music work within a sort of standard capitalist or market economy. Um, but you know, there's a lot of musical life that doesn't really fit within those contours. Uh, so it sort of straddles, um, you know, uh, capitalism and interesting and important ways. Um,

Now, you know, one way that this plays out with regard to streaming is that, you know, one way of understanding what streaming did was that it essentially didn't try to reinstate, you know, the sort of private good characteristics that music's digitization and its move online jeopardize and some people would say like fatally undermined, right?

So, you know, streaming at least initially was presented as a kind of alternative to the system of, you know, MP3 downloads that you could purchase, but more importantly, you know, the sort of rampant file sharing, which at least...

uh record industry representatives said were fatally undermining the the record industry as the kind of going concern in the the 2000s right the year of napster and then all the sort of um other file sharing sites that sort of grew up in its wake

You know, streaming, interestingly, you know, doesn't do what the record industry had done before, which was try to like turn digital music back into a private good by using things like digital rights management to like

you know, basically limit one's ability to say copy recordings. It didn't use like the threat lawsuits so that the record industry association America used in order to try to get people not to file share and things like that. Instead, you know, streaming platforms like Spotify just say, okay,

we're happy to treat music like a public good. What we're going to do though, is essentially limit the space within which it can be treated as such, right? So we're just going to put a fence around that, charge you a toll to access our platform. But once you're in there, it's basically an all-you-can-eat buffet. You can listen to as much music as you want. And, you know, they were quite explicit, you know, Daniel Ek, the CEO of Spotify very early on would say like, we want this to be

you know, very much like the experience of music piracy. We just want to be, you know, safe and legal, right? And, you know, have a sort of payment mechanism so that artists can be compensated. But, you know, essentially we want it to be like going to, say, the Pirate Bay where you could find any music you wanted and, like, have it instantaneously and stream it on demand. So...

One way of characterizing that, and this has been in an earlier chapter, I sort of say what streaming platforms effectively do is they decommodify music in relationship to users.

But that's just a way for them to then be able to commodify it in relation to rights holders. And here I'm borrowing from a Swedish researcher, Rasmus Fleischer, who wrote a short article where he made this great point, which is like, you never pay for music on Spotify or any other streaming platform. You pay for access to a service. But once you pay that access, the music's free. So it's effectively decommodified. So,

This is interesting because this changes the way we interact with music, I think, in important ways, right? And, you know, there's a rich, you know, literature, you know, not just exclusively Marxist, but, you know, obviously there are a lot of Marxist thinkers who've talked about it, like, what is the effects that, you know, the modification has on culture and on music more specifically, right?

um you know and go back and read adorno and see what he has to say about you know the sort of you know what you call the regression of listening and sort of the atomization uh um that he is sort of attributed to sort of commodity listening um but you know the question then arises well if music is being you know transformed in this way on streaming platforms does that then have um

you know, kind of comparable impacts on the way we engage with music as listeners, but also in the way that like, you know, musicians think about music, conceive music, produce music, perform music, et cetera, et cetera, et cetera. Now, one thing that, you know, I then argue in chapter five is that, you know, going back to something we were talking about before in terms of

You have the ways in which musical categories change partly in response to platform incentives to have music that's good for collecting data on mood, activity, context, things like that.

is that, you know, they incentivize people treating music in those terms, right, as a sort of means of adjusting one's mood, regulating one's mood, like accompanying various activities. And, you know, in a way, the sort of decommodification of music on streaming platforms really facilitates this because back in the days where you had to buy, you know,

CDs individually and it was, you know, quite expensive to buy a CD of like, you know, the standard price in the United States was $16 for a long time. You know, you bought the things that you, you know, you were going to listen to. And that had a very, that incentivized a certain kind of like engagement with music. Whereas, you know, if, if music's just sort of there, you know, you've paid your, your,

$5 a month if you're a student or $10 or $11 a month if you're an individual or whatever, if you're on a family plan or you're on freemium, you listen to advertising. It's just there, ready to hand, right? Like a utility plan.

right um then you know that really affords listening practices that treat music in a much more kind of promiscuous way to sort of accompany any and every activity right so if i'm doing work i can put on a focused playlist and that might help me concentrate if i'm working out i can put on like the beast mode playlist to help me get into the beast mode if i'm cleaning the house there are cleaning the house playlists um

And so, you know, the thing that's interesting is that, you know, essentially, you know, there are many ways we can sort of interpret this shift and this, you know, different set of kind of ways of engaging with music that, um, streaming is, uh, promoting, uh, quite actively. Um, but I think social reproduction theory, um, Marxist feminist, uh, uh, sort of Marxist feminist, uh, perspective is very helpful because essentially a lot of these activities that, um,

music streaming is promoting are, you know, uh, activities that are, um, bound up with processes of social reproduction, either at sort of individual level, right? Uh, how do I get through my day-to-day life so that I can be a productive worker and reproduce the conditions for the accumulation of capital by extension, um, which are, you know, also the, the, the sort of, you know, necessary conditions for me reproducing myself as a living being, right? Um,

But it also has those effects at intergenerational levels. So one of the things I talked about in chapter five is the proliferation of children-oriented playlists, but also specific Spotify kids apps and things like that, which basically use music as a tool for various kinds of...

Yeah. Childcare responsibilities. And then, you know, there's the sort of use of music as a kind of stopgap social service. Right. You know, one of the things I like to cite is this report that was put out by.

uh an all parliamentary group in the uk about seven years ago i think it's in 2017 that was basically documenting the sort of health benefits of the arts including music and essentially all this was uh you know extolling music for and the arts by extension for the ways in which you know they can sort of

improve people's health, their well-being, their mental state, et cetera, et cetera. You know, the sort of discourse that accompanies all this is like, look how much money we can save.

uh, on the NHS, um, on, you know, the, the sort of outlays that the government has to make in order to pay for things like elder care to pay for, uh, various kinds of healthcare services. If we can just use digital music as a sort of cheap alternative to that. Yeah. I mean, it was a couple of those examples really brought the, uh,

brought those principles from the abstract to the concrete, I suppose, with the music and the social reproduction. The kids' example really stuck out to me. My brothers recently had a baby and the extent to which

a particular song at a particular time can kind of be completely essential and avoid catastrophe. Maybe it really, yeah, it rendered the whole thing very concrete, which I, which I found fascinating. So, I mean, we've already gone reasonably long in the book covers a huge amount of ground. So I'd really urge listeners to seek it out because there's so much more in there that we haven't spoken about. I suppose you've already touched on it a couple of times, but yeah,

streaming and the digital economy in general is a very fast moving space. Are there any other big things that have changed since you finished the manuscript? Oh, wow. Yeah. I mean, I think there are a few things that, that it's not so much big things that have changed. I mean, I mentioned that the change in the revenue share, which is probably the biggest thing that's changed in the streaming ecosystem. I think that the, you know, it's more of what's changed around streaming. That's quite interesting. So I think one thing that streaming has enabled, which, you know,

It's not the only factor, but it's one major factor is, you know, the increasing acidization of music. So, you know, this has garnered a lot of attention in the media where, especially in the period around 2020, 2021, 2022, you had all these spectacular catalog sales. Initially, it was mainly older artists who were, you know, people like Bob Dylan, who's, you know, 80-odd and at this point in his life,

Holding onto his record catalog probably doesn't make much sense economically, and so essentially he sold it off to a record label in his case, or a major music publisher in his case, for a big payday that he can then...

do for estate planning purposes. But basically, you know, you've seen not just traditional music publishers and music labels, you know, buy these big catalogs, but increasingly boutique, you know, investment firms like hypnosis song funds, which has had a rough go of it the past couple of years. But, you know, there are others, primary wave, Harborview, investments,

And a lot of times these are backed by larger asset management funds or private equity funds. And basically the idea is that, you know, essentially streaming has set up a system where, you know, music is generating these small but consistent payouts and will do so for as long as the copyright term for

you know, these recordings persist for which, you know, and, you know, given the current state of copyright law in the United States, which is, I think,

pretty much the same throughout the world. You know, we're talking about, you know, upwards of a century after the artist's death, right? So, you know, these are good sort of long-term investments. They're also attractive to investors because they're what's known as uncorrelated assets, which is to say that people continue to listen to music even when like the business cycle takes a downturn. So there's that, you know, and that's something that sort of

exists in a symbiotic relationship to the streaming economy. It's sort of a development that's only conceivable once the streaming economy had matured to a certain point. At the same time, there's the growth of AI-generated music. There's an argument to be made that AI has long been a kind of feature of a lot of streaming services, at least at the level of automated recommendations and things like that.

um but you know definitely since 2023 this has become a major kind of cause for concern within the the record industry um you know the big thing being the fig drake song that was released uh this you know uh pseudonymous artist ghost rider upload a a song using like vocal cloning technologies in order to like

pass it off as a song by a collaboration by Drake and Moyle is just basically him, you know, using a kind of AI yeah. Vocal clone to sort of mimic the sound of Drake's voice. You know, so this was one of the things that has helped prompt the adoption of the artist centric payment method, right? Because there's a sense that this is going to become a,

A major issue for platforms, right, which already were, you know, the figures, who knows what it is, you know, as of December 2024. But, you know, like in recent months is, you know, something like over 120,000 tracks are being uploaded to Spotify a day. You know, if we start to see AI generated music take off, that number could balloon and you could get, you know, these...

services just being absolutely overwhelmed by you know what's ultimately this kind of garbage AI music and so you know I think that this is another thing that it's maybe not you know streaming per se but it's one of these sort of

adjacent technologies, which is going to have a major impact on, you know, the sort of shape that the streaming economy takes. I think, you know, the other thing to bear in mind is that, you know, it did look like in the United States, at least under the, the, the, the presidency of Biden, you know, and his FTC appointee, Lena Kahn, um,

There was going to be some regulatory movement in terms of breaking up big tech and changing some of the tax incentives that venture capitalists have benefited from.

With the election of Trump to a new term starting next January, given that he was backed by Silicon Valley, given that he was backed by a lot of crypto investors, I think the sort of regulatory environment is going to be much friendlier to large tech platforms, including streaming platforms. So who knows what that means for the long term, but it's something to keep an eye on.

and presumably even less friendly for your average musician, which tends to be the... Yeah. I think that's safe to say. Yeah, I mean, there was... I mean, I should also say there was a bill...

Oh, and now I'm forgetting the name of it, but there's a bill backed by Rashid Tlaib in the United States that was going to reform the streaming economy. I mean, chances were it wasn't probably going to make a lot of headway given that, you know, the Republicans had a majority in the U.S. House of Representatives last year. But, you know, the thing they would do was add like

a certain i think it's like a five dollar like supplement to every streaming subscription so if like you're a ten dollar a month like subscriber eleven dollar a month subscriber to spotify you'd now be paying sixteen dollars but that five dollars would be put into an escrow account then then distributed out to artists um

without going through labels or the platform, you know, sort of bypassing those intermediaries. So the payments would be made directly to artists in order to like, you know, up their earnings from streaming. But then also there would be like maximum incomes, right? So like somebody like Taylor Swift, you know, at a certain point, like I can't remember the figure off the top of my head, but like after like

million dollars or whatever, you know, she can't earn any more money from this revenue stream and it gets routed back to, you know, more sort of working class rank and file artists. Right. So it's sort of in many ways the reverse. It would have been the reverse of the artist centric payment model, which is

redistributing income upwards. This would have been redistributing income downwards. But I who knows, you know, maybe maybe maybe maybe this will pass, but I'm not going to hold out hold my breath in the interest rate.

And finally, what are you working on now? Right now, a couple of things. I mean, one thing that I, you know, I don't want to say I'm working on a new book just because I want to have like a little bit of freedom before I commit to anything. But I am sort of thinking about doing something more with the transformation of music into an asset class, because I think it raises all kinds of

interesting aesthetic implications, uh, some of which I sort of touch on very briefly in the second chapter of the book. But I think, you know, I think there, there's definitely like at least a book's worth of material there. Um, so maybe that's what I'll be doing next. Um, you know, I have an article that I'm working on on that. Um, and then, you know, I've, I'm doing some work on music and artificial intelligence. Um, and, uh,

And yeah, the big thing that's happening in the near future is I'm co-editing the Oxford Handbook of Protest Music with Norco Manabe at Indiana University. And that's currently in production, should be out at least online sometime next year, hopefully in print shortly thereafter. Amazing. Well, thanks so much for your time and look forward to speaking to you again soon. Yeah, thank you, Gumball.