What if Generative AI Grows the Music Business?

What if generative AI actually grows the music business instead of shrinking it? This week, Dmitri Vietze tackles the AI question everyone’s tired of hearing about, but from a different angle. 

First, Matt Adell, co-founder of Musical.ai, breaks down the $20 billion opportunity sitting inside AI music, and what it actually takes to get consent, attribution and compensation right instead of scraping music without permission. 

Then Gaurav Sharma, founder of Hook, makes the case for AI as an amplifier for fan engagement rather than a replacement for artists, and why even a fraction of social music consumption could be worth as much as streaming is today.

Stick around after the interview for a preview of next week’s episode with Daouda Leonard of CreateSafe, former creative director for Grimes, on the story behind Grimes AI voice licensing platform elf.tech and what they learned from building it.

This episode is part of the Rising Tide miniseries, working through the revenue multipliers that grow the value of the music industry, all leading up to the Music Tectonics Conference this October in Santa Monica.

Listen wherever you pod your casts:

Looking for Rock Paper Scanner, the newsletter of music tech news curated by the Rock Paper Scissors PR team? Subscribe here to get it in your inbox every Friday!

Episode Transcript

Machine transcribed

Matt Adell

[00:00:00] Dmitri: Hey, I'm here with Matt Adell. Matt is the co-founder and COO of Musical.ai, a platform helping artists and AI companies manage rights and attribution in the age of generative AI music. A veteran music executive, Matt's previously held leadership roles at Beatport, Napster, and Native Instruments, to name a few, where he consistently merged technology with artist empowerment.

And Matt has come to Music Tectonics, and he's a great person to talk about our topic today, what if generative AI grows the music business? Matt, welcome back to the show.

[00:00:38] Matt: Thanks, brother. I'm really happy to see you. and just, uh, one note, I love to say when people mention that I was a senior role at Napster, I like to say, "When it was legal, when we were paying people."

[00:00:49] Dmitri: That is true. That's a good-- a very good point. And I'm excited to talk to you because you've been on so many sides, the creator side, the streaming side, the B2B side, and now in this AI world, you're playing a really interesting role. And as we talk about the rising tide, growing the, value of music, I wanted to ask somebody who kinda had all those angles to help me figure out if this could be part of the grow the TAM.

If generative AI goes well, 'cause there's a lot of threats that it might not go well or it's not going well, but if it goes well, I wanted to, have somebody who could really help me paint a picture around this. So, let's dive in. What's the current state of generative AI, and how will it impact the music industry and revenue for the business?

[00:01:32] Matt: Well, I mean, that seems like, all the questions we're gonna try to cover in this short period of time. Let me try to answer, you know, the first one. We're past the is this thing real phase.

[00:01:41] Dmitri: Mm.

Mm-hmm.

[00:01:42] Matt: And, and now we're in the phase of developing an economic and technological infrastructure around it to make it work for the entire ecosystem.

 there are two major factors at play in terms of revenue around, AI products. The first is when someone subscribes to an AI service to generate music, how is that revenue shared with, rights holders and the people who create the music that the rights holders represent? And then there's a second, much larger, not yet talked about revenue pool, which is what happens when these outputs are out in the ecosystem generating revenue in the DSPs, in video games, in sync, in UGC platforms.

Uh, and that's what I'm excited to talk to you about today. You know, as you mentioned- Everything I've done has, in my career has been in some combinations of creation and consumption and monetization of music and protection of the artists who make it. And I'm sorry if you can hear my dog in the background.

I did lock her out of the room, but she's disappointed. and, this seems like the perfect place for me to be right now, because, this company and this effort to make this ecosystem work, uh, is really a culmination of everything I've worked on over the last, you know, 30-plus years

[00:03:05] Dmitri: Thanks, Matt. I think that helped us really dive into the picture there.

I mean, do we have any sense of how big this could be in terms of revenue for the industry?

[00:03:14] Matt: Well, a lot of people have different senses, but there's a PMP Strategy SESAC report from over a year ago that's really interesting that says that by 2028, the subscription to generative AI music services or the consumer fees, whether they're subscriptions or one-times or tokens-

[00:03:35] Dmitri: Mm-hmm

[00:03:35] Matt: will be a $4 billion a year business-

[00:03:37] Dmitri: Oh, wow

[00:03:38] Matt: ... uh, by 2028. And then the second pool, the, what we call the ecosystem revenue, what happens when the outputs are out doing business-like music, is s- projected to be 16 billion, these numbers are in euros by the way, by 2028.

[00:03:54] Dmitri: So you're saying that when you first described the overview of what the situation is, you said there's kind of like this upfront piece, and then there's-

[00:04:01] Matt: Mm-hmm

[00:04:01] Dmitri: this output piece. So that-

[00:04:03] Matt: Mm-hmm

[00:04:03] Dmitri: ... 4 billion that you mentioned would be kind of around the training and, people paying to generate these new tracks.

[00:04:10] Matt: Mm-hmm.

[00:04:11] Dmitri: And then this 16 billion would be once they release it out into the world, there's an increased amount of revenue that could come from that as well.

[00:04:17] Matt: Correct.

[00:04:18] Dmitri: Wow. So $20 billion is at stake. That's a, that's a nice chunk.

[00:04:22] Matt: It's, it's, it's substantial.

[00:04:23] Dmitri: And that could grow beyond that too, right? Like, that's just

[00:04:26] Matt: this report predicts, a doubling, in revenue every year for the foreseeable future.

[00:04:33] Dmitri: Okay. So now we've painted the picture of what the potential size is with growth going over time, et cetera, so forth.

Most of the industry conversation about Gen AI has been about the threat, the threat-

[00:04:42] Matt: Mm-hmm

[00:04:42] Dmitri: ... to artists and songwriters, the scraping of intellectual property without any kind of compensation, the lawsuits. But negotiations are underway and licensing are being discussed. We hear about things happening every day.

What would it take for generative AI to grow the music business the way this report is suggesting, as opposed to threaten copyright or threaten intellectual property? What, what needs to be there?

[00:05:04] Matt: Well, there's sort of a, chain reaction of consent and control. Rights holders and the artists and writers and performers and producers that they represent need to be able to consent and control their material.

There needs to be attribution that musical AI provides. There needs to be transparency, and there needs to be compensation. And I do wanna comment that the revenue generated by a service that allows consumers or music producers to pay to use a generative model, that's brand-new revenue The larger pool of what happens when the outputs are out on the ecosystem and people are listening to them and they're being used in sync and user-generated content.

Human beings only have so much time, and the number of human beings is a finite resource, so some of that is displaced revenue, not necessarily new revenue.

[00:05:57] Dmitri: Oh, interesting point. Interesting point. Yeah, because people who choose to listen to this might not be listening to something else that-

[00:06:04] Matt: Well-

[00:06:04] Dmitri: ... was already in the pool

[00:06:05] Matt: they're certainly not listening to it at the same time.

[00:06:08] Dmitri: Right. Right. And the rights holders could be different. The songwriters could be different. The, whoever owns that could be different, or they could be the same. I mean, I do wonder if there's... Maybe I'm getting ahead of myself in the conversation. I do wonder if there's a higher level of engagement if people who are listeners are also engaged with creating content, at the same time, the whole remix idea or, AI-assisted cover song, personalized music, et cetera.

[00:06:33] Matt: Well, there's no doubt in my mind that if you can give someone the tools to self-express, they engage more with the expressions they make and the expressions their peers make.

[00:06:46] Dmitri: Hmm. Right.

[00:06:47] Matt: There's no doubt.

[00:06:47] Dmitri: Yeah, that makes sense.

Yeah,

[00:06:48] Matt: yeah.

And I've worked in the digital audio, digital music creation space for many years, and that's always true.

DJs listen to more music than your mom-

[00:06:56] Dmitri: Hey, don't-

[00:06:57] Matt: ... unless she's a DJ.

[00:06:58] Dmitri: Don't be talking about my mom. So this kind of leads me to my next question. What are possible business models that make this a win-win for rights holders and artists, for labels and publishers, for fans and AI creators? what's the business model that gets us there?

[00:07:14] Matt: Well, we've already seen, the contours of the business model for how this is gonna work economically around the AI company offerings to users. Users pay a subscription or a one-time or a periodic fee to use these tools, and the makers of these tools, need to have a certain portion of that revenue go back to rights holders and then the constituents on, you know, who work with the rights holders.

 that's not completely dissimilar from, you know, a DSP like a Spotify or Apple Music. Certain percentage of revenue's going back into this pool and then dispersed to the proper people. And, that, revenue, needs to be based on licensed material. and we now see licenses being done, and there are AI companies retraining their models and getting ready to launch models that are fully licensing, license-compatible.

 and then what you need is you need attribution in order to determine if I've got this dollar I'm supposed to share back to the rights holder community, how much do I give to each rights holder, and then how is the rights holder divvied up among their musicians and composers and the publishers, obviously?

So, you know, we all know that at the-- in the world of the DSPs, there's a revenue pool that's supposed to go back to rights holders, and based on playback volume on a track-by-track basis, that money gets divvied up. You know, "Lucky Star" by Madonna gets played X number of times, and that's a percentage of total plays, so it gets that percentage of the money.

"TV Party" by Black Flag gets played X number of times, and it, gets its pro rata share of that money. That doesn't exist, for AI. And so what a- what we do with our attribution technology is we can measure the level of influence each song in the training set has on the outputs. And I think that's incredibly important, because There was a sense that, oh, we shouldn't even try to do attribution in this industry.

We should just do it based on volume. Uh, if there's 20 million tracks in the training set, we should just divvy up everything by one 20 millionth. And what that doesn't do is it doesn't value the individual value of the musicians and composers. And I forget the name of the track, but John Cage has a famous track that is just silence, right?

Let's say that's in the training data, and a Drake number one song is in the training data. based on the kind of music people wanna create, that Drake song is gonna have much greater value and influence and attribution in the outputs than that silent John Cage track. I'm using that as an example 'cause I don't wanna disparage any piece of music that might not be good.

 and so we measure that level of influence of e- the training data to every output so that it-- in the aggregate, you know how to share that dollar

[00:10:17] Dmitri: I'm now imagining people doing cover songs of that four minute 33 second John Cage song that using AI.

[00:10:23] Matt: Uh, people do. I worked on, when I ran, music at Wax Trax years ago, I believe Lead into Gold did a cover of that song.

[00:10:32] Dmitri: Yeah. But people are gonna use AI to do it.

[00:10:34] Matt: Yeah.

[00:10:34] Dmitri: And, and people are gonna download apps so they can just have silence. That's that's where it's gonna go.

[00:10:38] Matt: Yeah. That's, that's an

expensive and environmental way to create silence,

I suppose.

[00:10:43] Dmitri: Exactly. I guess that's what, uh, noise-canceling headphones are.

[00:10:45] Matt: Mm-hmm.

[00:10:46] Dmitri: But, yeah, I mean, I, I get it. Yeah, it makes a lot of sense that there would... There, there needs to be some way to evaluate/value which inputs are influencing a particular creation, not doing this sort of like massively equitable across just because somebody released a track somewhere someday, that it has the same value as something that actually was used in creating the prompt that created the song or, a cover song g- gets paid out to everybody in the world instead of for-

[00:11:12] Matt: Yeah

[00:11:12] Dmitri: the songwriters who actually wrote the song. So the... I see the attribution piece as being really... We're kind of at this critical juncture right now because that's still getting solved. what has to be in place for that attribution to occur now? I mean, this is your expertise, right?

[00:11:26] Matt: This market is developing not dissimilarly from how the user-generated content market developed, you know, 25 years ago, I guess.

 you know, we saw platforms, oh, just letting people upload whatever and play back whatever to other people. there were lawsuits and arguments about how that's supposed to work. And eventually, there were, not through a government mandate but through a combination of litigation, negotiation, and licensing, agreements that user-generated content platforms could use licensed music, and the vast majority of music got licensed to them, when...

and, but they had to install content ID Uh, into their systems as a requirement of the license, and use content ID then to figure out the payments downstream. The exact same as- thing is happening here. We'll see more and more rights holders, define their licenses as requiring generative AI music attribution.

 and then the AI companies, will work with us to do that attribution. Uh, we-

[00:12:35] Dmitri: Mm.

[00:12:35] Matt: Th- this is not completely new way of doing things. It's just a new technology in a very similar workflow.

[00:12:44] Dmitri: Yeah. Yeah. I mean, it's great to think that we might have a content ID system, especially an independent neutral one, not that's, you know, 'cause, you know, like with YouTube content ID, that was really the original model for what happened in other s- UGC video.

 it's owned by YouTube, and so, you know, they, they say, "Well, we can just take it down, or you could generate revenue from it." That's kind of a different conversation than saying, "Well, we'd like to have attribution from the beginning and negotiate based on the value of what's there," as opposed to, "Well, we'll just take you out of the equation."

So-

[00:13:14] Matt: Well, there is one really big difference between, say, a YouTube or a user-generated content platform and an AI company. Before the AI company can launch a successful product, they need copies of millions and millions of songs.

[00:13:30] Dmitri: Mm.

[00:13:30] Matt: and so they're not... whereas a UGC platform is just waiting for a user to upload a song.

[00:13:35] Dmitri: Right.

[00:13:36] Matt: Uh, and I'm no lawyer, but the DMCA, you know, gives YouTube and user-generated content platforms some protection against being responsible for what users do. Whereas what an AI company does, they can't wait for users to give them stuff, the stuff one at a time. They need to go get it actively in advance and maintain copies of it.

[00:13:58] Dmitri: Mm-hmm.

Mm. And apparently, they can't just use research training data sets that are out there that have been scraped either.

[00:14:04] Matt: Uh, research training sets meaning, uh, pirated content.

[00:14:08] Dmitri: Yeah, exactly. Exactly. Well, this has been great. I think really quickly, we've, already talked about what the potential value is here, what some of the challenges are, what needs to be in place, what kind of business models could come out of this.

I'm curious, once this is all said and done and we get all these things right, what shifts in music if we get it right? If we get the l- attribution, the licensing, the compensation, the opting in, what shifts in music? What does it look like?

[00:14:33] Matt: So I've always enjoyed playing with music. you know, from the earliest days of sampling, I make my own remixes and edits for when I'm out playing live, DJing.

 I understand that I shouldn't be putting those out commercially, and I don't. but I believe what we're gonna unlock is more, ability for human beings to express themselves through music. that's exciting. And, you know, that's the first piece of it. We are also gonna unleash massive volumes of material at a scale we've never, ever seen before.

 and that is gonna change how DSPs work. It's gonna change how consumption works. And I would imagine there are gonna be people who don't ever wanna listen to AI music. There are gonna be some people who don't care, and there's gonna be some inflection point where For the most part, you just don't know, which frankly has already happened in terms of AI-assisted creation.

You know, we've seen Tracksource and now Beatport say they're gonna block material that's 100% AI-generated. It's pretty easy to detect stuff that's 100% AI-generated. It's also kind of possible, in varying degrees, to identify stuff that's AI-assisted. but, you know, what no one's figured out how to identify yet is I use an AI service to create an output for me, and then I record a real human-recorded cover version of that output, which is an extremely common behavior already.

And at that point, there's no detector that can tell, right, you didn't write this, you know, that you got it, you know, from one of these outputs. so there's gonna be a real blurring of the lines for better or for worse. I think that's, uh, the for better or for worse is up to philosophy majors.

[00:16:28] Dmitri: Hmm. Yeah.

Yeah. So the DNA of the song came from a trained AI model, and there may or may not have been attribution at that point, but even if there was attribution at that point, then, the song's never been released publicly, and then a human band or ensemble-

[00:16:46] Matt: Mm-hmm

[00:16:46] Dmitri: ... uh, then records their version, human-created version, and so you don't really know who wrote the song is what you're saying.

[00:16:53] Matt: And that behavior is extremely common already.

[00:16:55] Dmitri: Yeah. Yeah. That's interesting because I was gonna ask you, like, if we get this right, talk to me about the user experience. You're, you're generating AI, generated music. Uh, I'm just curious, like right now, there's the text prompt version of create a fully finished piece.

Is there anything else that you think is gonna evolve even if we get this right? That's a great example of something that could happen there. Um-

[00:17:17] Matt: Sure ...

[00:17:17] Dmitri: anything else interesting there?

[00:17:18] Matt: I think ultimately, the write a text prompt and get a song out, gets boring pretty quickly as a user-

[00:17:25] Dmitri: Yeah

[00:17:25] Matt: ... experience.

So, I think what you're gonna see is integration of these capabilities in DAWs like Ableton and Pro Tools. You're gonna see the invention of new interfaces, that let you express yourself. I mean, let's face it, there are a lot of people out there who can't write a, great expressive paragraph, much less a sentence.

 but those people still wanna self-express, and so there's gonna be other ways, other tools that help someone do that. I think something else you're gonna see sooner rather than later is, recommended pre-generated assets based on your consumption history in the DSPs. You know, we're all really familiar with, "You listened to this, so you might also like..."

 you listened to this, so we generated this for you. We think you'll like it. Mm. So personalized listening. Is, is coming fast and furious.

[00:18:23] Dmitri: Yeah. I've heard people talk about personalized listening as a category.

[00:18:27] Matt: Yeah.

[00:18:27] Dmitri: And that's really personalized. Like, the song itself is, like, totally personalized.

It's not just, hey, it's a happy birthday song that says Matt Adell in it.

[00:18:34] Matt: Yeah.

[00:18:34] Dmitri: It's, it's

a totally different song.

[00:18:36] Matt: I mean, I don't know. I think functional music, exercise music, relaxation music, there's a lot of, uh, value probably in one-to-one personalization. but prime... You know, I think in most communities, music has a communal value.

 and so it, the recommendation will be personalized, but if there's a system where the consent allows, the system to generate a cover of Marvin Gaye's What's Going On in the sound and style of Wet Leg, you know, there's millions of people a DSP could recommend that track to. and, you know, survey after survey says that the reason people go to a particular festival is because their friends will be there and so people really enjoy listening to music with their friends and hearing and talking and sharing about music together.

So I, I do think the one-to-one music opportunity is, is limited. but I, I think that there's gonna be a tremendous amount of generated or generation collaboration material that people like listening to together.

[00:19:46] Dmitri: That's really interesting. It's fun to think about those more creative situations where it does increase engagement and increases number of hours spent with music.

It may not be listening to commercial tracks-

[00:19:58] Matt: Mm-hmm

[00:19:58] Dmitri: ... but it might be listening or creating commercial or music cr- based on commercial tracks that have been licensed, opted in, compensated-

[00:20:05] Matt: Yeah

[00:20:05] Dmitri: ... et cetera. So yeah.

[00:20:07] Matt: Absolutely.

[00:20:07] Dmitri: Well, it's so funny because whenever I talk about generative AI, I always confuse myself about the user experience or the musical artist's experience.

[00:20:16] Matt: Mm-hmm.

[00:20:16] Dmitri: And then whenever people start saying the word creator, I'm like, are you talking about social media video creators? Are you talking about songwriters and producers and performing artists? Like, they're all creators to some extent, and I know that most of the time people are talking about video creators, they're talking about social media creators.

But when we get into this UGC, personalized, hybrid, collaborative music-making experience, all of a sudden the lines get blurred really quickly. Mm-hmm. Yeah. So I do wanna ask about the flip side, and you've talked a little bit about the common use case that somebody might do a human-covered, cover song of an, a generative AI song that's...

They're using AI to help them come up with songwriting ideas, and then maybe they tweak it a little bit, and then they cover it or maybe they just cover it directly. And I could see, like, song camps or collaboration production workshops in studios in real time being like, "Oh, shit, we have studio time.

What are we gonna record?" You know? And be like, "Hey, quick, let's bring up one of these companies," and, "Okay, that's what we're gonna do, and I'll add a little this, and you add a little that," and before you know it, you've got a song. But what else about the experience for the musical artist, the people who already think of themselves, consider themselves artists now?

What are the positive use cases that you'll see once we get this right?

[00:21:31] Matt: Well, lots of artists, uh, and I'll say artists broadly and under that umbrella in this conversation, that is musicians, composers, lyric writers, and producers. and I generally wouldn't refer to them as creators although I do do it by accident sometimes.

And they're already using these tools, uh, in a lot of cases, not everybody, but way more than the average consumer listener knows today. and good tools for them that helps them self-express in a way that's efficient for them is, good news. But the first step of this is consent and control. if I represent the Prince estate or I am Wet Leg, in which case, uh, the two leaders of Wet Leg are also the composer lyricists, they deserve to be able to decide on a use case by use case basis how and if they want their artistry used by AI companies.

They deserve to know how that is used, transparently over time, and they deserve to get paid for it. I mean, what's the best use, case for a musician? To make more money, 'cause money is scarce

[00:22:48] Dmitri: Yeah, absolutely. And I can imagine, you know, w- we, we, people are looking at this through the problem lens because of the scraping that's already occurred, because the lack of attribution, and the lack of licensing, and the lack of compensation.

But I can imagine a place where, and we saw Grimes do this, um, but I think it was a bit early, where you say, 'Actually, let's train AI on my songs and give our users a chance to not only make new songs in my style, in my voice, maybe there's also the opportunity for the fan to generate some revenue along with the artist, too.'

[00:23:21] Matt: Mm-hmm.

[00:23:21] Dmitri: Or, you know, or, or to, kinda like a remix contest, too-

[00:23:24] Matt: Mm-hmm

[00:23:24] Dmitri: ... like a generative AI contest. It's like, "Hey, make my 10th song for my upcoming album. Here's the parts, and you add something. You pick the lyrics, or you-"

[00:23:35] Matt: Mm-hmm

[00:23:35] Dmitri: ... p- you know, "You mash me up with somebody else," or maybe it's two artists that collab.

I mean, people talk about-

[00:23:39] Matt: Mm

[00:23:40] Dmitri: ... collaboration is the best way to grow globally right now, especially across languages. You get an in somewhere else, and then you can grow your market elsewhere. So I could see some creative uses like that as well.

[00:23:50] Matt: Absolutely. I, I just saw recently that Beatport has or is running a, competition where you get the beats from their beat library, and, there's a famous house vocalist, and you can use a generative model to, use his voice but your song.

uh, that's gonna happen, and that kinda stuff's really exciting and interesting. And that artist, that vocalist approved it and is part of the program.

[00:24:16] Dmitri: Yeah. That's, that's super cool.

Um, it'll be so-

[00:24:19] Matt: I think with Grimes, there was a risk, which is if I let anybody, if anybody put their words in my mouth, it's only a matter of time before-

[00:24:29] Dmitri: Who

knows what I'm singing about

[00:24:30] Matt: they have me saying bad things that I-

[00:24:33] Dmitri: Yeah

[00:24:33] Matt: ... I wouldn't say. Um, then we're entering sort of the deep fake territory, so there need to be guardrails.

[00:24:38] Dmitri: Yeah. Yeah. I think that the contest setup where the artist or the management or the label get a chance to have, you know, where are the guardrails, you don't even ha- even have to spell them out, the just selection of curation stuff, I think-

[00:24:49] Matt: Sure

[00:24:49] Dmitri: I think makes a lot of sense. But, you know, there's some artists who'd probably be like, "Holy cow, I get to collaborate with my fans in a way that's, you know, these are the people that have put, boosted me up, who've made me, you know, really successful or-"

[00:25:01] Matt: Mm

[00:25:01] Dmitri: ... pushed me along or been the, the center of the word of mouth," could create a lot of cool opportunities.

So if-

[00:25:06] Matt: Look, it's been 35 years since the Beastie Boys made all of their a cappellas available for free on the internet for non-commercial use.

[00:25:14] Dmitri: Wow. Yeah. Super cool. So lots of potential creative things there. This has been great to dive in with you, Matt. You've done a great job of painting the picture for us.

Let's talk about your company before we go. How does Musical. AI fit into all of this?

[00:25:28] Matt: We are that attribution reporting layer. Uh, so if you compare MGC, machine-generated content, to user-generated content, we're the content ID layer who identifies what material is being used and, and to what levels it's been consumed and how it should be compensated.

We work with all kinds of rights holders, publishers, record labels big and small, collection societies, uh, sync companies, and we also, work with AI companies to enable the licenses they are doing, not that I'm doing, that, that they are doing to sort of glue those licenses together in a set of business practices and technology that make it easy and lightweight for the AI company to implement, easy for the rights holders to digest and then disperse revenue back to their producers, musicians, and composers.

[00:26:27] Dmitri: Wow, you make it sound easy by saying that it's easy. But I bet it's not easy for you guys.

[00:26:32] Matt: Uh, it's gotten easier.

[00:26:34] Dmitri: That's great. Yeah. Well, it's really cool to know that Musical Ai's out there playing this important role. It seems to be, like, a s- key missing piece of the puzzle that, a lot of the generative AI companies kind of, it's like an afterthought.

And so having somebody with your level of expertise come in and have the ability to do this, it's so critical to, artists and, all, the, whole musical community as well. And, you know, this is a great little, I think, introduction to a conversation we'll be having at Music Tectonics at the conference.

You're gonna be there. You're helping us put together a panel called What If AI Grows the TAM, the Total Addressable Market of Music, a New Licensing and Royalty Paradigm. So we'll be able to get into more detail on that panel. You'll have some other great panelists with you. We're still finalizing some of those.

And then you'll hang out with us on the beach by the pool, and people can ask all the questions they want and, and start to figure out how Musical AI can help them too. So I'm excited to see you there, Matt. You've always been a great supporter just by being there, bringing your-

[00:27:31] Matt: Oh

[00:27:31] Dmitri: ... positive self there and

[00:27:33] Matt: It's, it's one of my favorite conferences in the world.

 I love seeing you and your team and all the people who are there. regardless of where I am in my career and what I have going on, Music Tectonics is just not to be missed, Dmitri.

[00:27:49] Dmitri: I appreciate it. Well, I'm so glad you're gonna be there. This conversation's been great. You're helping us figure out the rising tide lifting all boats, and, uh, I'm looking forward to seeing you in Santa Monica.

Th- thanks so much, Matt.

Gaurav Sharma

[00:00:00] Dmitri: Music obsessed and artist first, Guarav Sharma is the founder and CEO of Hook, an innovative social music platform that enables fans to legally create and share remixes of popular songs on social media while ensuring, get this, fair compensation for rights holders. Under his leadership, Hook is redefining the intersection of music, technology, and social media engagement.

With a rich background in music technology and digital innovation, Gaurav has a proven track record of building platforms that align with the music industry's evolving needs. Thank goodness. Prior to founding Hook, he served as the chief operating officer at JioSaavn, India's largest music streaming platform.

What an honor to have you here. How you doing, Gaurav?

[00:00:44] Guarav: It's my pleasure. It's great to meet. Thank you for having me.

[00:00:46] Dmitri: Appreciate it. For those that don't know, what is Hook?

[00:00:50] Guarav: Uh, Hook is a social music platform. Uh, we are combining fan creation, consumption, and community into one place. we're the first platform where fans can do everything they're doing on social media today, expressing themselves with their favorite artists, with real songs, real music, fully licensed from day one.

We've licensed with many independent, as well as more recently, Universal Music Group as a major with more to come. if streaming music made music accessible, Hook is making music social, where creation and community drive consumption.

[00:01:27] Dmitri: Hot damn. So how has Hook evolved since you started it, and what does Hook look like right now?

[00:01:32] Guarav: Well, we started with a pretty simple thesis, that social music engagement is the next era. when you look at the amount of music being engaged with today on social media versus even streaming, social is dwarfing just the sum of music being consumed. Absolutely. And so early on, we really focused on creation tools, how to enable fans to create with music the way that they do today, remixing, but just even just creating UGC with music in unique ways.

And what we learned pretty quickly is that it couldn't just be about creation, that a community was really what was powering that consumption layer. And so today, we are the full loop. We are, the single place where you can create music with your favorite artists, share it across the ecosystem. We are not a walled garden.

You can share it across your favorite social platforms, et cetera. Um, we make sure that everything created on Hooke is per, the wants and needs of the artists. Artists have full control over it. and you can find all of the top songs on social media created by people like you and other fans on Hooke as well.

[00:02:39] Dmitri: Finally, I'm so glad to hear about a way to do this that is artist-friendly and, not in a walled garden, and is more like in the flow of how social media actually works-

[00:02:50] Guarav: Exactly

[00:02:50] Dmitri: ... as well as how music works, which is cool.

[00:02:52] Guarav: Exactly.

[00:02:53] Dmitri: the generative AI conversation in music, it's been a rollercoaster. From the outside, I'm not inside a company like yours, it's hard to tell what's going on and where things will land.

And at Music Tectonics-

[00:03:02] Guarav: Right

[00:03:02] Dmitri: ... we're focused on growing the value of music. If things go the right way, I'm curious from you, Gaurav, how will generative AI contribute to the growth of the music market?

[00:03:12] Guarav: hopefully quite meaningfully. I'll start by saying I think the rollercoaster comes from actually a core confusion around AI that replaces artists-

[00:03:23] Dmitri: Hmm

[00:03:23] Guarav: and AI that amplifies artists or amplifies fan interaction and participation.

[00:03:30] Dmitri: Oh, good point.

[00:03:30] Guarav: Right? I think those are very different types of businesses. and I think getting it right is utilizing it as a tool to amplify. Amplify the creativity of artists, amplify the interactivity for the fans.

 and when you think about the, value that it can provide at the end of the day, as I mentioned, social media 5X's the amount of music consumption on social versus streaming today. We're focused on consumers being able to express themselves through their favorite artists. and we believe even if we can monetize a fraction of that engagement, you're talking about, potential revenues as meaningful as streaming today.

[00:04:07] Dmitri: I love it. I love that you're thinking simultaneously, really holding in balance the fan, behavior, what happens on social media, what- how they wanna interact with artists and music, and then the artist's behavior, what incentivizes them to be engaged with fans in such a way so that there's really like a fair value proposition.

Both of them are getting w- what they want. Uh, I'm curious, what's your thinking about Hook's role in creativity and incentivizing artists as well as those AI creators? I guess y- you probably don't even think of them as... That's what, like, when you look at, like, a Suno or something like that, it's like, there's these new type of creator, there's this AI creator, but in Hook it feels like it's not the same thing.

But I guess before you jump in, how does this, all this stuff play into your thoughts on licensing deals? Because I think that's where the value proposition really plays out.

[00:04:52] Guarav: This is something I feel really strongly about. I don't believe that licensing is a tax on innovation, and I reject this idea that in order to be innovative, you need to, you know, like break the constructs in place to allow and enable artists to earn.

I, I don't think it's about AI versus artists. It's about sitting around the table and understanding the rights holders' needs, the publishers' needs, the sound recording rights, so the, recording rights holders' needs, the artists' needs, the fans' needs, and building, a business model that's equitable for everyone.

 but licensing is the price of admission in my mind, having built a streaming service as well. and from there you're able to iterate and work with artists and the rights holders to find a model that's successful. It took us... It takes time. It took us about a year and a half with Universal in, just for Universal alone.

 we went through around 30, plus 30 artist tests with the likes of The Weeknd, Lil Wayne, Lil Baby, Ayra Starr, like the list goes on. And we were testing what security requirements are needed for enabling this? What's the right business model? what sort of controls are required? I don't understand why it's that difficult to do that.

 it's not, and I really encourage everyone trying to innovate in music to just talk to the artists.

[00:06:11] Dmitri: I love it. I love it. I'm so excited that you're doing it the way you are. It's like you're explaining it almost like artists could treat this version of AI like they do streaming, like they do YouTube, like they do sync licensing, that it's another way to engage with fans, to increase audiences, and to monetize the use of their music.

How does that aspect, that whole engagement, growth, monetize, how does that play out with Hooke?

[00:06:35] Guarav: Well, on Hook concretely, an artist can have the, we distribute music into Hook in the same way a streaming service does. So we've set up our entire distribution channels in the exact same way, having come from streaming.

So artists put their music on Hook, and from there they can set their controls on how they want fans to be able to interact. We host challenges, on a weekly basis. In fact, right now there's a French Montana challenge on Hook where fans are remixing and creating UGC with his, new single, "Ever Since You Left Me."

 and from there there's a leaderboard, and then there's prizes, and the fa- and the artists themselves are able to nudge fans into certain genres of creation if they so feel. And every one of those versions that are created, you know, they're owned by that artist. As it's shared off-platform, we claim it for that artist.

So now one track is now, you know, all these different snippets of a new expression that that artist can monetize on, and becomes a multiplier of the asset as opposed to saying, "Let's take your asset and, as a song and, let's bifurcate that value." and so for us, that's where we see the potential as, AI as an augmenter.

 and for artists, they should view it as a way to boost their streaming revenues, their social revenues, all of that, through at least the way that we've, we're approaching it at Hook.

[00:07:53] Dmitri: I love it. I love the fact that this is not a walled garden approach, that this is really a-

[00:07:58] Guarav: It's

really important.

[00:07:58] Dmitri: Yeah. it's a multiplier. It's, an amplifier- Different ... out there in all the ways that, fans actually operate, a- the way that artists operate as well. It's not like it makes sense to build up a system that is like an alternative to a streaming service because nobody's g-

[00:08:13] Guarav: Right

[00:08:13] Dmitri: nobody's really gonna go there, especially if it's all, you know, fan fiction or, you know, like, unapproved remixes or whatever the heck it is.

[00:08:20] Guarav: Right.

[00:08:21] Dmitri: Or especially artists, why would artists send their fans there if it isn't even approved or licensed, et cetera?

[00:08:27] Guarav: Exactly.

[00:08:27] Dmitri: So Gaurav, I have the billion-dollar, maybe it's the trillion-dollar question.

How big could generative AI music increase the value of the music industry if we get it right?

[00:08:37] Guarav: Well, if you at least think about it through the social lens that Hook is approaching it with, music streaming, recorded music as an industry is a $30 billion industry, right? And as I mentioned a couple times on the call, social music consumption is almost 5X streaming consumption today.

UMG just reported for the first time in their earnings report that subscription growth slowed, and talked about short-form consumption as being one of those cannibalizing forces. So I do believe that, if it's 5X consumption, even if we can capture, you know, a fifth of that, right, we're talking about, contributing the same amount of revenue that streaming is contributing today, driven by AI.

So, that's kind of how I think of it.

[00:09:22] Dmitri: Amazing. This has been an absolute blast getting to dive in with you so quickly with Hook, understanding how you fit into the, our story of the rising tide, and the potential that you just described for doubling the music industry's revenue and, value as a result of this.

Guarav, it's been a pleasure. Thank you.

[00:09:37] Guarav: Thank you so much. Really appreciate it.

Daouda Leonard- Grimes

[00:00:00] Dmitri: Daouda Leonard is a philosopher at the forefront of the creator empowerment movement. With a decade of service in the entertainment industry, Dauda has worked as a manager, publisher, p- producer, A&R, and creative director for Grimes, DJ Snake, Skrillex, Bloodpop, and more. His creative contri- co- his creative contributions have generated over 15 billion global media impressions and sales to date for cultural leaders including Lady Gaga, Drake, Madonna, Ms.

Lauryn Hill, The Weeknd, and Justin Bieber. Dauda is currently the founder and CEO of CreateSafe Inc., a music technology company that provides SaaS and premium services for music artists, managers, and record labels. Dauda, welcome to Music Tectonics.

[00:00:44] Daouda: Hey, Dmitri. Thanks for having me.

[00:00:46] Dmitri: Yeah, thanks for joining.

So many people in the music business like to say that distribution has become commoditized, and there are a lot of distributors now, and we'll get into how they're trying to differentiate as well as add more value for their artists and labels. But let's acknowledge that they've opened the door for a lot of independent artists and artist-driven labels to build their businesses and enter the market.

I'm curious from you, Dauda, what roles have distributors played in growing the overall market of music?

[00:01:15] Daouda: Well, I th- like, I've been in the music business now for, like, 23 years and, um, up until about, I would say, like, probably, like, 2009, you know, distribution was something that was for people who already were, like, selling records out the trunk of their car.

Um, they were already spending money producing goods and selling them and they needed a distributor and, you know, they were sort of, like, hot in the streets and those types of deals weren't really, like, what the average, you know, artist trying to, like, get into the music business w- you know, was, were, were looking for, um, asking for, um, negotiating for.

Um, but around two- I think it was, like, 2008, 2009 is when TuneCore came on the scene and, um, I mean, like, look, I, I used to deal with companies like Koch, um, and, um, Lanspeed. Those were the first distribution companies that I ever worked with, and they did allow me and my partner Clinton at the time to, like, do independent music, you know, business.

Um, putting, you know, making albums and putting them out, um, without sort of, like, the need to get, like, some fancy record label to give us a deal. And that

[00:02:28] Dmitri: was primary phys- You're talking about physical at that point, right? Koch was known for that.

[00:02:32] Daouda: That was physical and it was also digital to a certain degree because that was when, you know, I think, and I don't know, I don't remember the year iTunes, um, launched, but that, we, we were doing a little bit of digital dis- digital distribution by way of, um, iTunes.

Right. Okay. Got it. Um, but it was primarily physical, like CDs, uh, vinyl, et cetera. Yeah. And so 2008, 2009, TuneCore comes around and it was the first time that I could just, like, we, you know, we could make a record and just upload it to the internet and get it on iTunes, um, and start selling your music. And that did open up the floodgates.

It was like that along with, like, the blog era, um, of music and SoundCloud. Um, the, you know, I th- I don't think sometimes people talk about bl- the blog era and SoundCloud as, like, sort of like the underground digital distribution that actually broke up... I, I, I think that, like, widened the pie. Like, clearly people were not really making money, um, distributing their music in their ways, but they were distri- they were finding and developing audience.

[00:03:53] Dmitri: Mm-hmm.

[00:03:53] Daouda: Um, so it was like the f- you know, a form of free digital distribution. And, um, so I think, like, what that did was it gave everybody now, you know, almost like s- you know, I guess like 16 to 17 years later, the ability to, um, get their music to stores.

[00:04:15] Dmitri: Mm-hmm.

[00:04:16] Daouda: Um, and I think that's, I, I think it's important that people really understand that's what really distribution means today.

Uh, like, when people, you know, think that they're getting, uh, a distribution deal, I think a lot of times people think that they are, um, now going to be marketed to the masses. Mm-hmm. That they're now going to get A&R services, that they're now gonna get to work with, like, other creatives and help them, like, build their career, when distribution is really just a service that gets your music from your computer to the computers of Spotify, Apple, YouTube, et cetera, et cetera.

Um, the meta platforms, uh, the t- the ByteDance platforms, um, and it's just getting you on the shelf.

[00:05:08] Dmitri: Right.

[00:05:08] Daouda: Um, so it's giving people the ability to get their music on the shelf quickly and fastly at extremely cheap cost. Like, it could be nine bucks a month, $100 a year or something like that 5%, 10%, 15, t- all the way up to, like, you know, 50%, depending on what you're negotiating for on top of just getting your music to these stores.

[00:05:36] Dmitri: So let me ask you something. That, I think that's a really great point, and we haven't brought it up on the podcast or on this miniseries yet, that it's really, it's not necessarily that distributors are growing the market. The first thing they're really doing is just creating the on-ramp, just allowing independent artists to even participate.

[00:05:53] Daouda: That's what it

[00:05:53] Dmitri: is. Do you, do you think... And you mentioned, like, well, SoundCloud blog era kinda led to, that was the first on-ramp, in a sense, of sorta like just put your music up, basically- Yes ... and get people to hear it or hear about it on the blogs. Do you think that, uh... And, and then you sort of implied, like, you know, sever- several years later, severalteen years later, now that that music, that pathway is there for totally independent artists to, to have their stuff in the stores, on the stores, the streaming services, the social media platforms, et cetera, now there's an opportunity to monetize.

I guess what I'm saying is, do you think independent distributors pave the path for a growth of the whole pie because of that pathway? Or was it people are listening to less commercial pop music and more independent grassroots music?

[00:06:38] Daouda: No, like, like, like, yeah, like th- w- for sure distributors paid the, paved the path for the pie to grow.

[00:06:45] Dmitri: Okay, cool.

[00:06:45] Daouda: Like it's, that's a, like, like that's a sheer... It's like, um, if you think about two, 2013 through 2016, um, you have DistroKid, TuneCore, um, let's call, let's, let's throw CD Baby in there, and let's throw even Stem in there at a certain point in time, like dur- like, like I, I think they're like 2015, '16.

But l- like, like let's just say now you have like four or five platforms that you could just sign up for and put your music out at a time period where streaming becomes the dominant way in which people are consuming music. And you have two, you have, you have two platforms that help you make the money, which is Spotify and Apple Music, and then you've got two other platforms that sort of like get you wide audience, which is like SoundCloud and YouTube.

They don't, at the time, they're not really like, you know, helping you make a lot of money in that sense. YouTube, to a certain degree, if you're like putting out videos that were like generating revenue, um, or like lot- lots of views. And like these, these intersected at a moment where like those platforms were growing in terms of people using them, and then now you have these platforms that are allap- e- enabling anyone to get their music there, right?

Just as fast and as quickly as it takes RCA Records or Interscope Records to get their music there. And that's a comp- like that's a moment in time where like- It's the, the, the level, the, the, the, the, the, the playing field was almost leveled. Um, and I, and I'll... And an example of that is, like, 2015 I started a project that was called With You.

It was, like, me and two, like, uh, three other producer friends. And, um, we put out a record called Ghost featuring Vince Staples, and it was sort of like this electronic hip hop sounding record. And I also, I just happened to know certain people who were working at Apple Music at the time, and I also was, had a, had a, had a friendship with this guy Justin Lobliner at Darkroom Records.

He ended up finding Billie Eilish later and breaking her as an artist. But it was like he understood how to, like, really, like, get things into the sort of like hype machine blog zeitgeist and, like, get it in front of people's faces. And my buddy Sean Glass really liked the record. He was over at Apple Music and he's, and he, um, socialized it to enough people in the Apple Music buil- you know, building that, like, we got this, like, premium placement with Apple Music.

Like something that would've normally been reserved for a record label Um, and we also got this, like, um, shot in the arm in the sort of like blogosphere. Um, so it was like right conditions that like nowadays, right now people are trying to look for these types of what you would call distribution marketing moments- Yeah

that are few and far in between. But because of the platforms that existed, like I was using a service called Foundation at the time, which I still, I still work with this company, Foundation. Um, they're one of the premier top independent distribution companies. You can't sign up for them. It's kinda like if you know, you know.

Um, um, and, and they were a, they are like a premium partner. Like if you go on Spotify's, I think, I don't... You could probably find this website today. Um, there's a, there's a website on, there's a, there's a page on Spotify's, um, platform that shows you these preferred distribution partners. And what that really just means is you are delivering what they call professional high-quality, um, music to the services.

And, and, and, and, and to dig a little deeper, what that means is your metadata is properly being delivered, like the files, like the f- like, like how you actually deliver that data to Spotify or Apple, et cetera, they look at that. When it comes in their system, they're like, "Is this, is this, um, organized properly?

Did they, did they get, um, like a certain amount of data that we require, um, and make sure that it's like delivered, um, legibly in, in some ways?" Like, that's the best way to put it. Yeah. Um, and, and, and, and, and Foundation is/one of those, and, you know, like ONErpm, you, you can look, like d- uh, like DistroKid is on.

It t- there's tons of companies that are on there, but that's where you can go and you can see like, okay, who does Spotify or Apple consider to be like the, the, the, the, the top preferred services of delivering the content to them? Um- And following their so, quote-unquote, "rules." Um, yeah, so all that to say, like, all these conditions got met at, uh, uh, between 2013 and 2016, where anyone could get online and start making, building an audience and making money with that audience, um, because the...

Like, not everybody wa- You know, like when Spotify first came around, not everybody understood it. Not everybody was, like, jumping to it. Not everybody, you know, was developing relationships with the artists and resident services. Like, I had a good friend, his name was DA Wallach. He was a artist, um, he was a part of a band called Chester French.

A- you know, post that band, he was one of the people who helped bring Spotify to the US and he created the Artists in Residence program, which is now what we call Spotify. It's the whole ecosystem of what we see as Spotify for artists and labels and manager services and all that stuff. And they had to go around and convince people to wanna get their music on Spotify.

So if you were one of those people that were convinced, then you were getting on those playlists. Um, you were getting that, that inventory that now is like, you know, now it's hyper-competitive. Yeah.

[00:12:53] Dmitri: Yeah.

[00:12:54] Daouda: But I will say from 2013 to 2018, it was like the gold rush of mu- you know, of music distribution and streaming, in my opinion.

[00:13:05] Dmitri: I wanna ask you about that because you're really talking about there's this moment where, where it's, the on-ramp's being created. There's kind of this parallel universe of, of SoundCloud and, and blogs where discovery or a free kind of version of some of the distribution was happening. Um, but then to get into, like, the actual stores where monetization could happen, you, you'd partner with distributors.

There's a lot of different tiers of distributors. But now it's changed again. You've mentioned the, the, the playlisting side of it, how that's become really competitive. So for a while, distributors were really pitching hard, like, "We can get you on playlists. We can get you on playlists." It seems like that's gotten so competitive.

There's been another evolution. I'm curious, what's changed in the recent couple of years around distribution? How do distributors differentiate now that there's so many options? What are they pitching to artists as, "Yeah, we're not just a distributor. Anybody can distribute you. Now we blankety blank"? I'm curious what your thoughts are on what, what's changed and, and how they're differentiating.

[00:14:02] Daouda: I think that point blank, they are all beginning to offer, let's call it artists and lab- like, like, like, like... Uh, what's the best way to put it? Marketing and promotions and creative services, right? Those are the th- like, the creative services are like the A&R/brand development, like money to help you, um, create your brand to make your music, et cetera.

That's the financing for that. Um, financing/people who can, um, help you with marketing and promotions. Two different things, you know, like marketing is like when you create a bunch of marketing assets and you decide to like send them to a bunch of places, whether that's like Instagram on your own accounts or other like, um, uh, people in your, in your, in your, in your sort of like, uh, m- management team, record label team to get to, to, to get it to as many people as possible.

And then promotion is like when you're on these other platforms like a radio station or you are using influencer mar- you know, like, like content creators for them to go and then push your music to their audiences or whatever. Sl- two slightly different things, but anyways, um, they are beginning to offer those as services as a part of a deal to, um, use their service versus another service.

[00:15:37] Dmitri: Right.

[00:15:38] Daouda: Um, but, like, I would say quite frankly that it is a myth still. I believe that, um, uh, unless you are doing a deal where a distribution company is spending enough capital that they then also need to provide you with, like, team re- human resources, right? Whether that's in the market that you're already based in or globally, um, you don't really get those services from any of these distribution companies.

They're all really just, like, technology companies. They're just really, like, they get your music from point A to point B, they collect the mon- they, like, like, where those services monetize it, they collect that money and then they pay you, and sometimes they have, like, payment splitting technology where you don't have to go about paying, um, all your collaborators and stuff like that, which is also a service.

Payment splitting technology, um, like most of, uh, some of these distribution companies do not offer that. Um, and then that is a pr- that's a service that you have to do if you have to pay for samples, if you have to pay for featured artists, you have to pay for producers, et cetera. Um, if you sell physical goods and you have to pay the publishing on that, like, that's a full service that you need to figure out, like, who's gonna do that job.

Um, and so I think that the, the thing that people are pushing towards now, United Masters has sort of begun to offer that, where you can, um, you get these, like, like, artist partnerships deals, um, which is technically just calling it a record... It's like, it's a record deal, but maybe without taking, um, your cop- your copyright in perpetuity, um, or, um, or a large piece, a larger p- you know, like-

[00:17:31] Dmitri: So, so you're s- so you're saying, like, the, the evolution a lot looks like part of a record deal, but not like owning your masters, uh, for a period or, or your copyright, but, but, but, but more like, uh, we'll give you advances and then we'll collect more until you recoup on that advance, but you'll still own your copyright.

Is that sort of the main way that things are changing with some distributors, with artists that they feel they can invest in in advance?

[00:17:57] Daouda: I think so. I think, like... And, and, and what I mean by that is, like, TuneCore Believe, for instance, like, they're stepped- they're stepping into that space. Um, as I just explained, UnitedMasters, I think that companies like, um, Too Lost, um, Empire, um, but Empire even more so some- sometimes feels more like a record label service business than a, just a distribution company.

But they are starting to, like, scale and begin to offer their distribution technology to record labels per se and say, "Hey," like, "just use our distribution infrastructure as a record label and you, and you go and do your job as a record label." Mm-hmm. Um, I think that is also what is the, the, the new trend.

Most of the, like, whether it's, like, Virgin, The Orchard or ADA, um, or Concord, what they're, they're looking to do is partner with executives, artist executives/executives and managers, people who develop the talent. And so they wanna give those types of companies deals, um, so they can go and build up catalogs and keep them within their, you know, uh, like, uh, market share Um, so those are the kind of the, the, the, the, the, the trends I'm seeing.

Like mo- most people who are like managers that I know, um, artists, most of the time people are like, "Yo, there's no distributor that is like actually doing more than just sending your music from point A to point B." Um,

[00:19:33] Dmitri: and- They have more services now ...

[00:19:35] Daouda: yeah, like, like, because it is really hard to service from, on a human level, the amount of people who are putting their music out.

I'm not saying that they don't want to or potentially do this with some of the high touch clients. I think it's more so you have to s- like we're at the age where some of the new distribution marketing artist labels, whatever you wanna call it, partnerships, technology is, is starting to be built. Like that's the era that we're in.

It has not been developed completely yet, but I think with the advent of, um, uh, coding, uh, agents, agentic, um, you know, uh, product development, you... we are starting to see that happen. Um, we are starting to see companies, like a, a company that I really like is Co-Brand. Um, when you like look at what Co-Brand is building, it starts as like sort of like a, um, a marketing platform, a way to like connect with content creators, see how your music is being used on Meta and ByteDance platforms, um, do direct to fan capture and also distribute your music.

So it's like when you think about that, like the direct to fan capture and, and communication part is A part, as, as, as a marketing service. You know, someone would normally do that for you. Same thing with, like, paying for content creators or doing digital ad buys. I think they do digital ad buys on there.

Like, those are three different services that a record label would typically offer now productized.

[00:21:21] Dmitri: Hmm. Yeah. Right? Yeah. And I guess that's sort of where I, I see some... You mentioned, you mentioned, like, the funding, the advances piece of it with distribution, but I'm hearing a lot of distributors talking about little, other little marketing service, whether it's content creation, like video-related things or, um, or mark- marketing technology to help them expand the, the reach of their audience, segment their audience, target their audience to get the link in bio type stuff to drive people towards monetization as well.

So I'm seeing, like, there's the companies like you're describing, CoBrand, that are independent... Well, I was gonna say they're independent of distribution, but you said they also have a music distribution- They do have distribution ... component- Yeah, yeah, yeah ... as well. But they're starting from a marketing perspective.

There's a lot of interesting music tech platforms emerging that handle, maybe it's, uh, primary data leads capture for fans, maybe it's, uh, CRM, managing your emails and mobile addresses of fans, whether it's for recorded music or live performance and ticketing, and then helping you segment and reach them and engage them in other ways.

We've had this company OpenStage on the podcast before that does that kind of thing, or the, like the Lelos that have the drops or the Symphony OS that have the advertising, uh, f- support for, for artists, the, the Rivets, things like that. So there's these independent companies, and then some of the distributor, distributors are han- are, like, m- either doing deals with those companies and merging that as kind of like the artist services or the marketing tech services th- things like that as well.

Um-

[00:22:46] Daouda: I, I would bet money that all of those companies either become distribution companies, if they're not already, or they get bought by distribution or major label groups to be a, a, a, a, a, a, a, a, the technology aspect of that, that they can then offer that as a, as a service as to why you sign to their distribution.

But I don't see them being like-

[00:23:11] Dmitri: Is, is that because it's just crazy to have like, "Oh, this is my distribution partner, this is my this kind of marketing, this kind of marketing, this kind of monetization"? Is that, is that why? Is it just too much friction or is it because there's more money if you're connected to the distributor and ultimately they'll have to do distribution in order to, in order to survive?

[00:23:28] Daouda: I think it's more friction for the end user who's a, who's a artist sometimes. Like, why would you wanna... Why, why, like, at some point in time you just wanna make music and put it out and release it- Yeah ... and, and, and focus on that, the, those creative aspects of your business. Um, and you w- you want to work with a company that can do a lot of those things for you as, as, as you become more successful.

Um, I think that there are independent entrepreneurs, like the people who wanna build record labels, who are maybe more open to using like, like, like building a stack, um, that they can then be like, "Okay, I'm using Symphony, I'm using, um, uh, EVEN, and I'm using Too Lost." Let's just, just random, "And that's my stack."

Um, but like I think as a, as a, as a, as a business, you have to sort of have that mindset and, um, and also skillset. Like it's like learning, like a lot of these tools, like I don't care how, how simple everybody thinks even AI is nowadays, there's a learning curve to a lot of this stuff when it comes to like operating at a high level.

Um, you know, there's a difference between like ChatGPT free and Codex. Like this is two very different learning curves and, um... And I say that to say same thing with like a lot of the music technology platforms that exist today. Um, I think that they have been built for, um, executives and not necessarily artists, and why should they be like a, like...

I just, it's a very weird world where you're asking artists to do all this like work work, and they just wanna make, they wanna like do things that pertain to like- Be, you know, expressing themselves, making the art. Maybe like the, the, there's a, there is a, there's a artistic and creative aspect of marketing that they also wanna participate in and are totally down to participate in, going out and promoting their record and different things like that, and then doing shows, um, engaging with fans, et cetera, et cetera.

They, like, asking them to do spreadsheet-style work is just, you know, it j- it just feels like that's not what people sign up for when they sign up to be an artist. Um- Yeah. Yeah,

[00:25:52] Dmitri: gotcha. Yeah,

[00:25:52] Daouda: yeah ... managers and marketing directors and project managers and et cetera do, and it's like, all right, like what stacks can you build to build out a, an, you know, a music business ecosystem that supports an artist in a major way?

[00:26:06] Dmitri: Yeah. Speaking of stacks, a lot of what people come to our Music Tech Talents Conference to do is to build their tech stack for music. You mentioned CoBrand. Real quick, are there two other music tech companies or services that you found valuable for artists? Let's do just a quick speed round on a couple, 'cause I really wanna get to ask you about your Grimes AI project before we run out of time.

But, but let's start with those two. Are there two other services or music tech companies that, that you think have done well for artists?

[00:26:35] Daouda: I really like the product Untitled. Hmm, what do they do? The Untitled is a, um, place for storing, um, music that has not been released. And, uh, like for many years, like most people, like in my position where like you're always looking for like a way to listen to music, store the music, um, I don't like, like I don't necessarily...

I'm not using Untitled for the purposes of like storing the files of the music, but it is, it has the, has one of the cleanest interfaces closest to like, let's say like iTunes or Apple Music, um, in terms of like it feels like a music listening product and specifically for unreleased music

[00:27:21] Dmitri: So, like to share, share among collaborators and colleagues and pitching internally in the industry, not to consumers?

[00:27:27] Daouda: Yes, yes. Nice. And, and- All right ... um, but they have functionality where you can order vinyl, like you can, you can, like, you can, like, you can, like, now you can go from being unreleased to released on their platform. So it's a- Wow ... it, it's a, it's a, it's a... They're, they're building some of the functionality- Of a Bandcamp

for pre-release marketing. Yeah, yeah. Like- Damn ... like a, like a new type of Bandcamp. So-

[00:27:49] Dmitri: Interesting.

[00:27:50] Daouda: Okay ... you know, like watch what they do in the near future.

[00:27:53] Dmitri: Nice. That would... One more. Do you have one more music tech company that's helps artists, especially around like revenue growth or monetization, uh, artist services, things like that?

[00:28:03] Daouda: Um, and I can't promote my own company, right? Well,

[00:28:06] Dmitri: you can. Yeah. Tell me, tell me.

[00:28:08] Daouda: I mean, look- Let's do

[00:28:09] Dmitri: it ...

[00:28:09] Daouda: we, we, you know, CreateSafe, like technically we are, um, we have distribution as a service. Um, but what, but, but basically what CreateSafe is, is, um, a metadata management company. Mm. We help you take all of the data you have about your business and leverage it.

Um, and an example of that could be, um, taking all the information about your songs and figuring out, like what, what are unregistered songs, and- Mm ... those songs that need to get registered so that you can collect that revenue that's just being generated because you've put out songs on DistroKid and TuneCore, et cetera, but you haven't registered them with the MLC or PRO, et cetera, um, or SoundExchange.

Um, another example of how we, our, um, how our tools are used are to do like complex, uh, f- financial forecasting, like figuring out how much money you need to spend on marketing, on, um, on the, the, the production of the music. Um, so budgeting, expensing, uh, long-term forecasting, catalog valuation. Um, so like, yeah, music metadata is like gold and oil.

Um, and so it's all the files, you know, your royalty statements, your song splits, you know, information about your collaborators, uh, your audience, you know, your d- your, your fan, um, data, putting that all in one place in order to leverage it and build, and, and, and build a, a, a plan around how you're going to, um, run your music business.

[00:29:42] Dmitri: Wow. I'm, actually glad you explained that because when I read about CreateSafe in the past, it was associated with the Grimes AI project. Can I ask you real quick about that since we do have a conversation going about, how generative AI can lead to more money in the industry, but mind if I ask you about that one?

[00:29:58] Daouda: Yeah. So look, we, you know, in 2023, this was coming off of our initial, seed round fundraising to build, you know, a new distribution platform, off of what we had built in terms of metadata management. And the idea was that, like, p- new platforms, new products, and new ways to monetize music are gonna exist, and what are gonna be those new ways to distribute your metadata into these new platforms, products, et cetera.

 mid 2023, around, like, April, is when the AI boom started to happen, and it started to happen in music around people cloning voices and making, you know, songs with other people's voices. And what we did is we built, um, Grimes ended up, seeing sort of what was happening around, like, people cloning Drake and The Weeknd.

And where other people were saying, "No, I don't want anyone to use my voice, my name, image, likeness," Grimes, posted a tweet that said, "Anyone who wants to use my voice, feel free as long as you compensate me 50/50." Similar to how most, like, recording featured artist agreements are done. Like, someone wants to, like, have you on their song, most times the artist or a record label will charge 50% of the net profit that comes off of that master recording And so she was just sort of putting what, what insiders know out there saying, "Hey, if you wanna do this, you can do this with my voice."

But, um, there were all these voice cloning platforms that were out there, but there was no way to actually get the music onto the platforms, split that revenue properly, and ensure that Grimes got paid and the people who were making music with her voice would get paid. And so we took one of our initial distribution products and direct-to-fan capture products, and we combined them to create what we called the music game at the time, called elfTECH.

And what elfTECH was, was know, you sign in, it looked like her computer, almost like where you go and make music, and you upload your vocal, you turn it into Grimes' vocal, um, you make your song, you upload it back to the elfTECH product with clean metadata. You, you know, the same way like, like you would use any distribution service.

And we distributed to Spotify, Apple, et cetera, and we split the revenue, , with the fans who used her voice. And, um, we did a partnership with TuneCore as well, so that anyone who used TuneCore as a distribution service could use T- TuneCore and split the revenue back to, you know, to, to Grimes as well.

[00:32:26] Dmitri: I mean, I love the model because, like, right now the conversation is generative AI companies that have not licensed music, they're scraping music. There is no model in place for attribution or paying artists back. But this was one where it was artist-driven and it could potentially create a whole other revenue stream for an artist.

So I'm curious, what did you learn about the potential for that revenue stream with the Grimes, experiment?

[00:32:50] Daouda: Well, what, I think what we learned is that, like, fans just wanna engage with their favorite artists. Like, like I don't think that anybody cared that much that it was AI per se.

What they cared about was engaging with their favorite artists. what we also learned is that, like, we're very early in this sort of like AI cycle, so like there's still to this day not a lot of people who want to do this, so it was like a very one-off sort of idea. and the last, I think the most important thing that we learned is that technology can be used to create, or to, augment how we do old, inefficient processes like featured artist clearances or licensing data.

Like, essentially what we built was a robust backend that anybody can go in and, on, you know, on Team Grimes or whomever is using this type of product and, approve songs for distribution. they can review them. they could respond back with like alternative terms. this is just sync licensing.

This is the thing that typically happens in a day-to-day business where you, like someone's trying to figure out how to sample someone's song, how to collaborate with them, and if you have this front-end interface that's linked to your website, that's your brand's web- you know, like your artist's name, image, likeness, you can give people the opportunity to sample your music.

 and it's as simple as that. Like, you don't, like, yes, AI can be a part of it, but this is now just opening up the licensing pie-

[00:34:25] Dmitri: Yeah

[00:34:25] Daouda: ... in general. Um-

[00:34:27] Dmitri: I

[00:34:27] Daouda: love it

... and so, and AI will be, it is, can still be a part of it, right? Like you can still, through this system that we built, you can license your music, and data to an AI model, right?

 like that is something that people are not talking about. Like- the paradigm that we're in, there should be technology, which we've built a version of that and we're gonna continue building. There should be technology that allows artists to control their name, image, likeness, and data associated that on any platform.

They should be able to say, "Hey, here's my S3 bucket." Right? They don't have to, they don't have to know that language. We'll come up with new language around it. But- An

[00:35:08] Dmitri: AWS server, basically .

[00:35:09] Daouda: Yeah. Here's my AWS server lo- you know, API credentials, MCP, that has all of the data and instructions on how to use my data.

 The same way you have a DistroKid account or an Instagram account or a et cetera type of account, you should have a CreateSafe metadata account that allows people to access your data with certain permissions.

[00:35:32] Dmitri: See, what I, what I like about this is It comes from the artist rather than coming from the AI company.

You know what I mean? It's, it's saying, like, "This is my creative world. This is my name, image, likeness. This is my style. This is the world that I created musically and all the things around it. I now give you permission to use it in certain ways. We're gonna share in the revenue," versus, "Hey, the Internet's a playground.

Let me get everything I can. Look, everybody likes playing with everything. It's like Instagram for music," except you haven't asked the people that you've trained everything on, and so all of a sudden it's... So it's really, it feels much more like artist permission. it's not just opt-in. It's, like, really collaborative at the licensing level.

And it's... I'm with you. I feel like if Beyoncé and Taylor Swift are doing what Grimes does and they say it's kinda like a remix or a cover song contest, but it's not with musicians, it's with their fans. Their fans don't have to say, "Oh, I need to learn how to play the guitar to do a cover song."

Instead, they're just, like, logging into an app or something, and there's like, "I've always wished that she did a song like this. I'm gonna do it," you know?

[00:36:37] Daouda: Totally.

[00:36:37] Dmitri: "With her voice."

[00:36:39] Daouda: Totally.

[00:36:39] Dmitri: You know? So I, I think this is... We have a whole... I might have to take this portion of this conversation and plop it into our AI licensing as a-

as a revenue optimizer, 'cause I, I think, uh, I think we're onto something there. Uh, I appreciate you sharing that. Let me ask you one more question, and then I'll let you go. This has been super fascinating. We started someplace different. I, I'm so happy we've been having this conversation. Here at Music Tectonics, we're pushing forward this idea that everyone in the music industry should start rowing in the same direction to get more people paying for music, whether it's this interactive engagement stuff we've been talking about, or for new uses of music, new types of licenses.

Dawda, what else should we be thinking about to grow the value of the music industry? I'd love to get your, your hot take on that before we wrap things up.

[00:37:25] Daouda: Well, I, I think, look, I started out in this business doing direct to fan. Um, I was, you know, the first thing I did was use Constant Contact to help build the email list.

Um, and I think that as primitive as, uh, CRM technology and stuff could be, I think that, like, that's where the, the, the, the... Everybody should be rowing in that direction, which is like, um, like figuring out how to do fan o- fan engagement and fan o- you know, ownership of your fan data in a more efficient way, um, instead of just relying on, you know, these, these, uh, platforms that set the rates for y- for how you're gonna, your, your music is monetized.

I think that, like, those are, that's like the radio, but, like, people learning how to sell music to their fans, their audience, to different partners, I think is the biggest opportunity right now. Like, we should be teaching classes on how to sell music to anyone, right? Like, the average person doesn't know how to s- to ask, "Hey, I wanna get paid a dollar.

I wanna get paid $2 for X, Y, Z." Um, I think, like, sales, you know, and marketing classes, um, you know, should be something that, like, distribution companies and record labels and even publishing companies should be teaching, management companies, like, what is the way to sell music? Because for the last 10 years, we've definitely been taught, like, give your music away for free.

Um, and then the second thing is, like, put your music on these platforms and let these platforms sell your music for however much they wanna sell it, versus, wait, when did we stop learning how to sell our art in general? Like, that's the job.

[00:39:22] Dmitri: Yeah. Love it. Love it. That's perfect. Dauda, it's been an absolute pleasure.

Thank you so much for sharing your thoughts on Music Tectonics. You've really contributed to this rising tide conversation. We've got upcoming episodes that will cover things that you just brought up, direct to fan and e-commerce, marketing tech that helps you segment and collect your first primary data so that you can reach out to folks.

But it's been an absolute pleasure. Thank you so much for joining me.

[00:39:44] Daouda: Thank you. Take care.

Let us know what you think! Find us on LinkedIn, and Instagram, or connect with podcast host Dmitri Vietze on LinkedIn.

The Music Tectonics podcast goes beneath the surface of the music industry to explore how technology is changing the way business gets done. Weekly episodes include interviews with music tech movers & shakers, deep dives into seismic shifts, and more.