The Future of Music Discovery Through Context & Emotion

Speaker A

Is, you know, this enormous amount of music that's available now, you know, that that people have access to everything, but they often don't know what to to listen to next. And the playlisting technology that's in some of the platforms is actually fairly limited. It's pretty pretty old school and primarily drives things based on po popularity. So Music Maps really there to bridge the gap with that, to basically bring the context that you're experiencing the music, who you are at that moment and what you're trying to do. With the choices that you might have in front of you. This podcast is produced by Amplitude Ventures, your partner in early stage startups.

Speaker B

Today on Sound Connections, Ty Roberts and Gorin Anderson join us to explore why music discovery still feels broken despite unlimited access, how context and emotion could reshape recommendation, and what the next generation of music technology needs to understand about listeners. Okay guys, I have two legends in the studio. Ty, Gurin, welcome.

Speaker C

Thank you. Great to be here.

Speaker A

Yeah. Great to be here.

Speaker B

we're gonna talk about a lot of things today and I'm gonna really satisfy my my nerdy side. so and and I have a lot to speak with with both of you. but we're gonna sort of really start the conversation around Moosey Map and your involvement. but we'll we'll get to that and we'll get to a lot of nerdy stuff. So I hope everyone is is looking forward to this. So maybe Ty if you could begin. who are you and what do you do?

Speaker A

I what do I do? Ty Roberts. I'm the founder of Grace Note. I created the company that basically helped make metadata a standard across both music and film and television. And I sold that business. I founded it and I sold it to it's owned by Nielsen now. And I left there in twenty sixteen and I went from there to be the CTO of Universal Music Group, which was very exciting right at the time that social media was really getting going. and so I was there during that time and worked on high-res audio. And long ago in the long past, I'll mention that I was David Bowie's collaborator on multimedia projects and computers. I helped David do a bunch a bunch of different things with music and the internet. And so I've got some history with artists such as

Speaker B

Yeah.

Speaker C

Yeah, so I'm I started as a recording engineer producer back in the 70s. And sort of born out of a n sort of natural curiosity, I used to love to discover things and found was very comfortable being in recording studios with creative people. I consider myself a creative person, but I really enjoy being sort of the if it's a band before musicians, I've actually like the f the fifth band member, but not really a band member. And then hopefully if I did my thing right, something really good would be even better. And that led also to quite a bit of other things in ex in exploring music around the world, I started to realize, you know, how do I find things that I know I will love? When I finally find it. But that sort of journey leads on to why I got really excited when I met the Muslim Map founders. And this was remember, this was a long time ago when I started this company. It probably started like 20 years ago with this mission to sort of map the music of the world with genres, moods, and music intelligence, as we call it, to then be able to really. find the stuff that we know we're gonna love if we finally find it. And now when we have so much content being released every day, I think this mission is even more important than ever.

Speaker B

Hmm. So wait, it's gonna be really interesting and and we sort of need to set the scene a bit. So we just need to describe music map a tiny bit more just to sort of understand the context. So Ty, you can probably describe this a lot better than me, but sort of in my interpretation, the Music Map and especially these new products you're launching, is really trying to understand emotional, contextual, and cultural dimensions of how music is organized. So basically, how can you categorize music and how can you find it? but maybe in a bit more precise manner, like how would you dissect what museum app does and and why are we even talking today? Like what's the topic?

Speaker A

Right. So Music Map has actually a long history. Music Map has around for actually quite a long time and it started with musicologists who basically were building a human edited database about music, realizing that there wasn't really descriptors that described the genres and the emotional feeling and what was really going on inside the music, which would lead a human to want to enjoy it. It wasn't just the factual information, like knowing that the track was from nineteen eighty nine was not necessarily the thing that's going to bring you to the music. It's more like, what's the emotional impact? What what what what are the lyrics about? What language is it in? Where is it from? What is the style? And so they built a very elaborate hand-edited database, I think about a million tracks, where they hand annotated all this information for a broad set of music. And that's the basis of the ground truth set that MusicMap has. And then on top of that, they built audio analysis technology. Which basically goes into the tracks and looks at what's happening and registers the beat per minute and comes up with different kinds of sonic attributes, which essentially allow further tagging of the musical track from a I'll call it a sonic perspective and an emotional perspective. And so that's really where the technology has continued to expand over the last decade or more to the point now where we have extremely accurate descriptors for the you know, the music itself, and we have the ability to relate that music to the humans who are Listening to it as well as to the use cases, be it in a car or be it on the phone or be it at the gym, you know, the ability to tie those things together. And that really leads to the discovery experience, which is, okay, I'm at the gym or I am in my car and this is what's happening. What should I listen to next? And as Joran has said, there is, you know, this enormous amount of music that's available now, you know, that that people have access to everything, but they often don't know what to to listen to next. And the playlisting technology it's in some of the platforms is actually fairly limited. It's pretty pretty old school and primarily drives things based on pop popularity. So, you know, while it's because it's popular doesn't mean that you're gonna like it. And doesn't mean it fits a situation you're in. So Music Map's really there to bridge the gap with that, to basically bring the context that you're experiencing the music, who you are at that moment and what you're trying to do with the choices that you might have in front of you.

Speaker C

And I think to add to that, because you know I've always been out looking at these things. I was involved in Echonest a little bit before they were forced to sell to to Spartifactors that were running out of money. but also turned out to think good for everybody. But what was really amazing with Pierre and the team at Musima, they had a strict editorial focus. I also met the Pandora guys and they were also doing similar things. But there was no structure, there was no focus. And and I remember, what's his name Vesterman? We had a big heated discussion. Actually he loved it. Because he said, so you have all these crazy people out there. So yeah, we love them. And I said, but there's also a lot of fans that might be giving you data that is not necessarily sort of correct, or if they might be just inflating things. And he said, Yes, we also have another problem. All the all the fans That hate other bands and then they put a lot of crap into the into the genome, right? So sorry, Tim, if you're listening. but you did say this. and I think that was what impressed me with Muslim app. They were very structured. It was going into like an editorial team at the magazine, where they met and they said, this is what we're gonna stick to, and because then you can have Like the grand truth we have is solid. And I think that's no one who has done this. And that I think is very impressive, something we're very proud of.

Speaker B

You know, it's really interesting because w for me when it's you know, it sounds like someone started twenty years ago out of passion and interest or maybe academia, but it doesn't sound like it's obvious this would turn into a a business twenty years ago when people started you doing this manual labor. Like w what what was the incentive or what was the foresight of the commercial value twenty years ago when they started undertaking this tie? Do you understand? Do you know

Speaker A

Well, yeah, yeah, yeah. So first of all, what it chang obviously when there were CDs and that was the business with selling albums and CDs, understanding all the music on the album wasn't actually as important for people. They're there the reality is is that they had to kinda like buy into the whole album was pretty expensive. They had to feel like there was enough music and popularity was probably a lot more important. But once iTunes came and you could buy individual tracks Once the music was fragmented into the ability to buy a single single, every album became singles. Then you had to pay 99 cents for a song. So it wasn't like streaming now where you have access for free. You had to pay 99 cents, but you only make a commitment to one song. So recommendation and these kind of things became really important. And and Apple, of course, introduced their own services for that, but so did the other platforms. So having the ability to get to the tracks that you were paying for, that was the first motivator. And then later, once streaming came and you had a subscription service. Then the problem became get me to the stuff that I like because I have everything, but I don't know what I want. And Grace Node did some studies of people at that time. And sadly to say, the average person knew about 15 artists. If you said, name me artists, and you made them sit there and name it, they would come up with like 15 names. So obviously there were quite a bit many more hundreds, tho hundreds of thousands or millions of artists than that. So the problem is that that text search. didn't really work well enough to get you to to what would happen is they could start with something that they did know and say, I want more things like this. And so so, okay, well what are more things like this? And you can do it down a factual line of information. But what Musimap has would do it down really a sonic and emotional line to try to really get you something that's basically accomplishing us a cult, an emotional journey for you on the music. And you know, that can be that can be very sophisticated the, you know, the playlisting today. doesn't necessarily create a musical arc of your journey. So, you know, like if you think about it, you start here at this place, we're gonna have these raised peaks, we're gonna get very excited, then we're gonna fall off and get more relaxed, and then we're gonna go back and get more excited, and then we're gonna have the big finale. This is how music is organized in film and then other things like that. And today we probably don't really think about a musical journey. But this is becoming super important. And in fact, some of the work we're doing with automotive manufacturers right now, they have a journey. You're going from work to home. You're going from home to soccer practice picking your kids. So, and it's exactly this long. And so what am I going to do in that timeframe to to help me really enjoy the drive? And and so they have a real journey. And I think that's actually turning out to be true for people that are traveling or for people who are exercising. It's a good one because you're going to exercise for an hour and you need a complete musical journey. It's kind of like, Heat up and get ready and then cool down. Like you need all that stuff in an exercise program. So we have the right data at Music Map to power such as such an experience.

Speaker B

Karen, maybe you can comment. As a music producer that's created so much amazing music, and we were just discussing before the call, some of the the Danish legendaries that I have referenced to. but the complexity of understanding, curating, describing music from a technology perspective, you know, it's probably quite difficult and I would assume also not all encompassing. But do you believe like we can actually through technology create something that really is taste based creation, emotion based creation that's really effective?

Speaker C

I truly think so. But also there's the even more basic thing that I see happening with younger people today, that actually they are creating albums with the playlists. actually no and one of my old friends in Stockholm, who was also at a record company, and he said he tried for so long to get his kids to appreciate his vinyl collection. They didn't care about it. But then he said, wait a minute, what are you doing? So they were actually creating they were recreating those albums on playlists. And without even so they were they were so here we think we can we don't just think we know, we can then give tools for people to actually build these collections. Like create soundtracks for your day or whatever, in a much more meaningful way than what you now get from the streaming services. so I think this is and we are agnostic, we can we work across all platforms. So this is tools for for music fans to really then dig deep into exploring collections and and then seeing where we also have relationships between the musicians that play across different records. And things like this becomes then an interesting discovery. And I agree with you, Ty, like people don't remember. I don't even remember. I I was involved in probably recording about a hundred albums. Even if I tried hard, I couldn't even remember them. Right. So of course I do remember them, but if someone asked me to write down the list in chronological order, I would be s struggling. So I even someone that works with music, we need this to keep our things organized and be able to find things. So This is sort of part of the fabric of culture. And one

Speaker A

If you talk about the car the car, you know, in car, people cannot people are not supposed to touch their cell phones anymore in the car. Like it's against the law in most countries now. So you're left then with voice commands, which then they would need to remember the name of a song. They might not have, we just mentioned the artist name. So you're really disabled there. It's not really able to in the car. And by the way, it's the number one place where people listen to music, is in the car. So so how do we get them? to the right stuff. And so having these other parameters that we can derive from a seed song, like they might listen to something on radio and they go, I'd like more like this. And then it's going to jump off radio to a streaming service and try to generate something that matches what their mood. And we would have, I think, a pretty good shot at that. I do think there's a lot of other factors, you know, which one that one when one builds a complete music experience, it's like already knowing what their favorite songs are. knowing where they are located, knowing what's happening around them, knowing the kind of taste profile they have, with how adventuresome they are with music. So there's many other factors that go into to playlisting than just purely the emotional information. It's it's in of super important, but there are other components.

Speaker B

So how widely spread is technology like this? you know, I can only speak for myself. I I only use Spotify. I've I've done so since it wasn't even open in Denmark unless I you know borrowed a Swedish account from one of my friends. and you know, I've been a big user of that, but but my biggest pain point with with streaming services is I just can't find playlists that I can enjoy. Like I run a lot. I I run many, many hours a week. And for me, listening to music is just a pain in the ass because I I just can't hone in on one vibe, which for like me like running is like meditation. So it's my bigger frustration with any streaming. So how far have they come in being able to create these emotional contexts rather than other alternatives that I don't really even fully understand?

Speaker A

So you're a runner, so the problem is you run and you're it's working and then you have to stop and like find another song. Is that like what actually happened?

Speaker B

Yeah, but because you know, for me it it it's my way No

Speaker A

It's not not really a good thing with running.

Speaker B

no, it's my way of meditation. So I I wanna stay in a certain mindset, I wanna stay in a certain rhythm, I wanna stay in a certain mood, and I just can't figure it out with these traditional streaming services. But try to describe to me like what are they doing and then what what might you be be doing better?

Speaker A

So Yorin would probably talk about so Echo Nest, when was that company around and when was that purchase? What year is that?

Speaker C

That was Christ, goes back to early days of MediaNet or Median Lab, right? It was the year that SoundCloud got their award. And then a couple years later sold to Spotify. I mean not SoundCloud, but but they couldn't.

Speaker A

Two thousand ten, maybe, two thousand and eight. Somewhere in that time frame. So so the reality is is that the first generation versions of this technology were really around that time frame. And to solve some of this problem. And Apple had their own version of this, but probably metadata data driven, you know, music genius is what they called it. And my company, GraceNote, provided the data for that. They had some very simple sonic analysis like beat per minute. They had some very basic stuff, but they didn't have any of this advanced emotional data. And as we move forward in time, these systems haven't really evolved. And that's why we think there's an opportunity for this for our company and then certainly the reaction we're getting from the marketplace to rethink about this because actually what's happened is the amount of new music is like accelerating. And of course there's a whole separate topic we'll talk about later AI music. But the reality is people are getting buried in the total number of tracks that are out there. And And they feel lost, a lot like what you're saying when you're running or they're in the car or they just don't and it's a it's a problem because if they haven't if they don't a aren't able to find the thing that they enjoy, then they might choose to do something else. And so we really for enjoyment of music, it's super important that like people find what find something they like and get into the vibe, as you mentioned it, and and enjoy it.

Speaker B

Mm. But so so you know, it's it's hard if I play devil's advocate on that statement, it's hard for me to imagine that these big tech giants haven't come too far or haven't evolved a lot. Like what what is the reality of this space right now when it comes to sort of discovery, recommendation, and algorithms?

Speaker C

Unfortunately I think these streaming platforms that you know we do depend on and I think they are they brought business the business back to generating money again. but they are What they do as far as playlisting is kind of self-serving. They are in they're in the business of programming music that then generates revenue for the industry, but not necessarily generating excitement for the users. And I think that's and we we feel we can add an extra value to the fans. So one thing that's been sort of my North Star in all my work. There are great artists. Arguably it's it's up for discussion if they're great or not, but they're artists that create stuff. and then there are fans that love it. Everything else sort of between those two bookends is subject to change. And that's also where we sort of find ourselves working. So I feel I sort of I serve two masters, the artists and the fans. And and if we can make that experience Better, like when I'm in the studio working with an artist. If I can make that recording to be, well, that was really good. you know, it's a good way to end a session, right? Walk out the studio with that feeling. and then the fans get a good kick out of it. I think another thing that you you sort of hinted at when the founders at Muslim Lab started this thing, when they're thinking about How to turn it into commercial success? Probably not. But then in my view, that's actually a good thing. Like all the records that I worked on, of course I wanted to be successful. But we were there to create the greatest record each time, right? And of course, we all expected them to be successful. But if you're only sort of focusing on How this is going to be commercially successful within the studio, you're probably setting yourself up for trouble. Same thing when you create a new venture. It has to have a real purpose for the customer that will be using it. And then if you succeed with that, I've I'm I'm in the belief it will be successful, also commercially. And I think, you know, with all the music tech startups. The ones that keep that as the North Star, they become successful. If you just look at a presentation where they're talking about sort of their exit plan when they haven't even built anything, that's kind of when I say, I don't know if I want to be involved with this. So I think if you if you really serve a purpose in the creation of your art or your business, then you have a good chance. And and I think we, you the It hasn't been easy to build these successful streaming platforms, but I think we have something we can add in an extra layer of discovery and and and excitement for the fans. I'm sure about it.

Speaker B

Because it it's, you know, I I'm I'm slowly unpacking my understanding of of the problem statement here in this industry. Like Ty, you would you would be better than most to comment this, but right now I'm working a lot on on data with ISWC and RSSCs for for a company that we've invested in. and I'm I'm definitely not a developer, but let's call me like a a very advanced vive coder. I've been building products for years. So and you know, I'm looking at sort of what does MLC have, what does other data provider have of codes and and it's just a complete mess of codes and inconsistencies and and matching that that's really, really difficult. So maybe Ty you can come into what is the status quo on the problem of the lack of data, the lack of matching, the lack of understanding of music, what consequences does it have for the industry right now, and maybe how does Muslimap play into that?

Speaker A

Yeah. I mean we can talk about the metadata, but I think the the first thing is we were talking about discovery and what I would just say about that is it hasn't evolved a lot. the one challenge you just have, just the most basic thing is when you wanna find something new, the tendency is to look at the data and play the most popular thing. So if you say, Dave Brubeck, it's gonna play take five. And if you're a music fan, you already know that song, okay? Same as you play You say, play the doors, and it's gonna go light my fire or play, you know, break on through. Like that's the first song you're gonna hear. Because they of course know that that's the most popular song for most people. They're not really understanding exactly who you are in the sense of like, you know, you're looking for something new, are you looking for something so the context of like when you approach the service is is kind of like the next dimension. Like what what am I trying to do? You're going for a run. You know, they probably could actually figure that out if they could look at your travel path and see that you're moving at, you know, not our speeds at human speeds. you know, down a road or on a track. And so the reality is is that that that context is is missing and the way to connect the context then to the music. And the current products don't really do that very well. That's why there's specialized exercise products. You know, like you you could get one of these specialized that understand exercise and the types of exercise better and have a warm up and a cool down period. And you know, Nike and others tried to make these kind of products. And so It's a specific problem for exercise. It's different for me being at home, different for me being in my car. So what I would just say is context awareness is super important and not that hasn't evolved a lot. And I was just gonna say in the basic music services, because they serve so many people, the product is fairly generic. There really hasn't been a lot of evolution in these products under the hood or otherwise, even if there's new data available, they're very similar to what they were ten years ago. And so Yes, maybe they've improved some of it, but but not lots. And and I don't know why that is. In the sense it's what they have is working, so I can't argue with that. People are using these products, but the overall streaming subscription business isn't growing very fast. So my argument would be there needs to be better product. Just like every other product, you know, would you people improve the products quite a bit. So that there does need to be some reinvestment in music and making it a better product and helping understand the music fan better, like or is telling. is probably a good idea. On the metadata side, which you brought up, yes, music is really complicated. So having all the publishing information, understanding who the songwriters are, understanding who the contributors are to music, this information is something that Musimap is actually assembling right now. We're building this kind of relationship because who's behind the music and what kind of things are in the music in terms of instrumentation or how it was produced are also key indicators as to what why people won't be interested in it. You know, a lot of people are interested in electronic music. There's no lyrics. So people are there are focused on who the producers are, who contributed to these tracks, who did the remixing, you know, so that's a c a very popular style of music that's not driven by the vocalist. And so and there are many, many examples like this where the data behind the music is important. And I think Musimap will assemble, because this is a part of its, you know, I'll call it genome. the understanding of who's behind the music in a bigger and broader way than anybody's ever done. And so that's something we're working on right now as part of our search, musy search product.

Speaker C

So these are the the things, Jacob, that I think, you know, we are we have several, you know, products. The these are some of the you would say we're gonna be geeky conversation, these are some of the more nerdy things that we do that maybe the music fans wouldn't really care about, but it these are the things that make the experience for the music fans a better one. And then as as Ty mentioned, you know, quite a f few years ago we worked on this to really look at the listener. So the the way he or she likes to what he or she likes to listen to. And and that that's something that I don't think anyone is doing. And then man those things together is a better way to have an have playlist generated or just searching for music in general.

Speaker A

Yeah, we have a we have a product called Musi Pro. So with you know, we we you know we have Musi Tag, which talked about, which is like the emotional tagging. We have Musi Search, which is basically is the searching a vast pool of information for something that you might want to find. It's being designed to work with voice search and things like that right now. we have Musi Profile, which is what we're talking about now. And so, you know, in that profiling, you know, you know, where we are, where we're gonna go. It's gonna become quite sophisticated because there's many different types of music consumers. I'll call it there's the casual music consumer who, by the way, was off served by radio. Literally, they just turn on radio channel X and listen to music. They just want music as a background to why they're driving or whatever. They weren't so focused on, this is the doors and this is a song about so and so, or this is, you know, Chapel Roan or whatever. They weren't focused on that. They were music was kind of like a backdrop to their life. So that casual consumer still represents a large number of people that are out. Then there's the adventuresome listener, you know, like the person who's always trying to learn more about new music, who wants to be the person who knows about music, the latest trends, what's happening. And so you can see from their musical taste that they're always expanding their knowledge. And they may be expanding their knowledge only in one genre. They're going to become the most knowledgeable person about what's happening in pop music today. Or maybe they're multi-genre. They're like, I like pop music. I'm going to learn about jazz. I'm going to learn about, you know, world music. I'm going to go into these other characteristics. So are they. How are they adventuresome in their musical taste? Then there's a listener who, you know, a lot of people really discovered a lot of music in their youth and in their, you know, high school and college era. And then they kind of stop their exploration of music and they only want to listen to that. You know, like, you know, my era, that is, you know, like be dating myself but saying like the talking heads. Okay, so I love the talking heads. I still love the talking heads. I know actually I know some of the members of the band now. But the reality is there are people who only listen to that era. And And so what they would be really interested in is when, you know, Jerry Harrison, who's one of the producers of Talking Heads, the members, releases a new live recording of old Talking Heads tracks from that period, they're all fixed up and fantastic. So that's a new Talking Heads thing they haven't heard before. They'd be interested that. So the problem is that in the world of music services today, I'm not sure that all these different contexts are really yet being considered. And we hope to empower that kind of next generation. Cause we think if we make the experience better for people. They become more in love with music. They learn more about music. They broaden their tastes. You know, you know, I feel like, you know, maybe people are trapped at the music buffet on the cruise ship. That's where they are right now. So the food that's at the buffet, but what's not at the buffet is sushi, what's not at the buffet is, you know, Korean food, what's not at the buffet. It turns out there's a whole broad world of music just like that. And we have to broaden their taste. You know, music is little bit of an acquired taste. Definitely is. Mm-hmm.

Speaker B

Can you break down for me how how this worked to some degree? Because it sounds like, you know, there must be a good amount of intelligence going into knowing how to understand a person. Like peop people you know, are probably by now pretty familiar with like, you know, chat GPT memory and and it clawed memory that sort of has these selective memory points that sort of remembers who you are and what you do. Is it somewhat similar? How does it work in its like most essential form?

Speaker A

It's music profiling is like profile in context. So this profile of this music in the car with just me in the car. Or if I have my, you know, I'm married, I have my wife in the car. And I know if I can give this example, I like b I like Black Sabbath. I like metal. Okay. My wife cannot listen to more than one second of that. Like literally almost get out of the car if I had that on. So and then she likes, you know, Joni Mitchell and Neil Young and this kind of stuff that's kind of like light. Californian classic rock, that's her thing. So when she gets in the car, we have a middle road which is called the Eagles, which is which is the least rocking thing that I can possibly tolerate, and the most rocking thing that she can tolerate. And so the profile would have to figure that out so that we can have a peaceful drive in the car and you know, maybe we'll be able to listen to some, you know, acoustic renditions of Black Sabbath songs or something. But the reality is is that it's gotta figure that out. And so There's profiles, there's context, and then there's profiles with multiple people. Cars are great environment for this because you know, when you have your cell phone, it's mostly about you. You're like in your pods, it's mostly about you. But when you're in an environment with other people, like you when you have party at your house or something like that, you have to think a little bit about, you know, like so my house, you know, the Black Sabbath has to go down unless it's, you know, a metal party at my house. And so I do think that that context profiling is a kind of a next generation. And then there's an exploration. Profile, which is really like, okay, so if I'm gonna go and learn about new music, in what dimension am I doing that? You know, is it decades? Is it genres? Is it locations? You know, like, hey, I know about the scene in San Francisco in the sixties. Do I want to find out about the scene in Seattle? You know, for grunge, do I want to find about the LA music scene? You know, what do I want to do? It could be location based. So people often travel, by the way, and then they go to the location. They try to listen to some of the music that in the place because they hear it around them, you know, when they're there. You go to I mean, y Yoran spends time in Brazil, so you know, when you're there, you're hearing Brazilian music. And so people suddenly discover that and then they want to learn more. So these are all the problems that people need to solve. It's just that, you know, they need t advanced technology to do that.

Speaker B

So is it you know, the the skeptic in me being listening to, you know, playlists you want to Spotify and and streaming services since they've sort of been digital are like that sounds like utopia. Like, can that actually be done?

Speaker A

Yes, it can be done. They do it today. I mean, they have this EchoDec and other technologies. They do when it's generating the recommendation, when it's not on a playlist and it's just like gotta figure out what to play next. So they're doing something to try to figure out what to play for you. and there's a lot of that by the way, you remember we're on the internet, so playback of songs aren't instant. So, you know, oftentimes this 'cause the b worst thing that could happens is you close the app. Like the next song and the next song want what you want and you go, forget this. I'm gonna go watch TikTok. Right. So that's like bad. Like if you're a music service, you don't want everyone to close the app. So oftentimes they will generate the next song, which they think logically based on what you're playing is the song you should hear. And there's a lot of factors that go in that. But actually in the background, they're buffering a guaranteed winner for you. Because if you skip the choice that they make, because you don't listen to a song you don't like very long, you might listen to it two seconds and go, I hate this. Click. They need to immediately play something you're gonna like. So usually they've built a few backups in their plan, which is something they already know that you like, not something new, something they're and they're gonna go to the guaranteed winner while they like recompute and go, okay, we didn't like that that choice, we have to figure it out. So you can imagine behind the scenes, there's a little wizard in there trying to keep you happy. And the question is, how how smart is that? And I think the different platforms approach this problem in different ways. And and you know, while we're talking about a sophisticated offering. I think that this is what will happen in the next generation of these products is that we'll become context aware, emotionally aware, multi-person aware, you know, these kind of things. I don't think today that they they do all of that, but they probably do some of it.

Speaker B

So Gorin, what what's your comment on this? Because as a music producer, is there any sort of resistance towards this intelligence, you know, taking so much of of that curators and sort of job to some degree or that part of of music?

Speaker C

No, I don't think so. I think this is part of if we if you look at what we what we call music profile, but it really the very basic thing you do, you basically get a playlist from that person, and that's been then the the basis of what we will serve up. And then we're now looking at having multiple people then because music is social, right? So you know it's It's a shame that most of the music now is listened to on headphones. So we also then have to take into account more than one person in the room. But I think this is actually we did this. I worked very early in my digital career on a project with Quincy Jones called Q Radio. That was quite 30 years ago, 28 years ago. It was very early exploration of. what music on the internet should be. It wasn't downloads, it was basically streaming radio from all over the world and having a record store. But fast forward to just some years ago, we made we asked Quincent to give us some of his favorite songs. And we generated then his profile. He's arguably very creative, was unfortunately he's no longer with us, producer. He was blown away. And and then he invested a little bit of money actually, music. and became really a spokesperson, an ambassador for for this vision. So I don't think at all people also we had one thing in common with me and Quincy. When people asked, what kind of music do you like? Then we knew the the the question was. To answer you know jazz, hip hop, whatever. And we had one thing in common, we love good music, and I still do. So good music, of course, highly subjective, but back to the idea that most people don't listen to just one genre. So people are multi-genres. And we are putting a lot of effort into with the profiling how to find this interesting sort of essence. of what I or anyone else consider to be great music across multiple genres. And I think that is probably our secret source. We are way ahead of anyone else in that space. So then we can serve up, you know, I love music soundtracks. There is this Vim Venders movie about a guy who cleans toilets in Tokyo. It's on one of the best soundtracks. Of course now I can't remember the name of the movie. But you can easily look it up. It has everything on that soundtrack. And it's just the greatest soundtrack ever. I wish I had produced that. And I think that's sort of what I think is missing. You know, we have these fantastic, yes, like I said, soundtracks from movies that sort of live up to this, I think, aim to really create. multi genre i music experience that just blows you away. And and and that's what I would like to, you know, that's one of the things that I I know we can do. And we are very you know, I think we are doing it actually. So so that's sort of what's missing. And and I don't know if that answered your question, but I feel there's no conflict here. The technology is it's just tools to me. You know, you know, I've been Around making records a long time. I remember when there was people saying, my god, synthesizers is putting, you know, string quartets out of business. And then I have string quartet friends say, Well, we're gonna use synthesizers and we can merge the two together. You know, I used a lot of sampling keyboards. you know, tools that are used by creative people are not the threat. The other things that are threatening us, but that's not it. So tech is It's a fantastic way to advance what we do as as you know, as creative humans.

Speaker B

Okay, so I have a question because it's you know, it's it's interesting that this is just not not only utopia, that is actually something that can be done and is being done. tie back to your comment about sort of I'll just mention Spotify doesn't need to be that platform in particular, about that changing a lot, and maybe incentives are not so to change it a lot, and maybe that's also unknown why they're not changing. But you know with that in mind, just because you have the best product out there does not necessarily mean that these incumbents want to use use the product or, you know, embedd the way that they do this. So therefore, the distribution or the impact of your technologies is sort of limited by distribution. What's your thought about that with commercialization and distribution? Like how how can you get this out to people?

Speaker A

I well, first of all, you know, the way I always approach this is there's always smaller players, there's always innovative people in the space or people who have specific problem. So for us, you know, the automotive industry has a very specific problem. They need a better musical interface that takes into account the fact that it's a shared music environment in a car, which is the opposite of the mobile application. So so that they're not getting that from these services. They're getting, you know an in-pocket experience, which is kind of what the mobile app is designed to do. And they're trying to adapt it to the living room on wheels. So this is a place where we can really do something. And also there's other issues like at a card or safety issues where we need to keep the driver awake. We need to keep the driver happy. We need to keep the passengers happy. We need the kid to keep the kids happy. So I think that that's a really great environment for us to work in. and then I think that on the service side, these guys are all competing with each other. There are multiple music services. And they have different focuses, like they actually see their market differently, the consumers that have them differently. So I do believe that you'll find use cases in there. So for example, you know, iTunes now is or Apple Music, sorry, is is focusing on live music. So that's an interesting problem area because there's l lots of live music, but it's not very well tracked or coordinated, or when you were mentioned before, the data that's available on it is not that great. So As you start to build a tune into live broadcast, that's one thing. Radio is part of these applications, you know, they're adding sports, they're adding podcasts. So, you know, there's a lot of non music content that's also in these products, some of which are podcasts about music, which I guess this kind of is. And so the reality is is that that actually all these things need to be organized because people have only so much time in their day. And so, you know, if mm it's one of the things that we believe is like for the future of a vehicle is what I'll call multimodal playback. So you get into your car and you see you're kind of a little sleepy and you know how you're going to work. And so it's gonna wake you up, then it's gonna give you some news, then it's gonna give you some entertainment, then it's gotta give you some sports, and then a little bit of Razmataz before you enter the office. And that that whole program is switching across all different kinds of sources. And so You can see that's not a that's not a function right now that's in the major music platforms. They kind of want to keep you on music, that if you switch to podcasts, it's fine, but they're not switching back and forth. But we're in a world where people switch back and forth, especially young people. So we'll we'll see how that how that works out. We think we're a great role for us in that.

Speaker B

So just to sort of understand it correctly, like an example could be sort of an interface layer over the platforms. So, you know, that could be like the way that I use Alexa in my kitchen to can you play this type of music? And it sort of understands the context of what that is and then chooses to write music on the platform. Do I understand correctly?

Speaker A

Yes, and that and you are exactly right, right. That example's perfect. The the overall platform in the vehicle is gonna be the voice. And so and so the problem with that is is that while these certain services have their own voice products, the vehicle manufacturers need to have a voice that understands all the features of the vehicle, like the air conditioning and the safety features and the engine. And so there is gonna be one big voice in the car that's gonna come from the auto manufacturer, probably. And then there could be some other voices that are part of the different tools or technologies. So you know you can turn on Claude the app inside of your car and then Claude voice is gonna come on or whatever. So yes, I think to some degree there's gonna be a little bit of an overriding voice system that's gonna basically derive the entertainment. And it's also connected to other systems in the car, like the navigation system and knows where your destinations are. It's also has the ability to monitor. Remember, people get in the car and the Bluetooth connects to their phone. So it kind of knows, you know, the kids got in the car because their phone's connected. Susie's phone is connected to the Bluetooth now. So anyways, I do feel like that kind of future is definitely coming. And that's a little bit beyond the single user mobile application function that's mostly what streaming music services are today.

Speaker B

Hmm. And that's obviously super contextual. and also touches upon how much information we want to share. What's your consideration considerations about that?

Speaker A

Well, there's obviously privacy information you need to allow if you want the system to know to help you, then you have to share what you need to share. It doesn't need to know more. But but the reality is is that so far people seem to be willing to share musical taste. You know, their personal identifier information, there's rules about that. But you know, the but the the reality is your mobile phone is kind of like an identifier for you, even if you don't know who's behind the mobile phone. You know, phone X is in the car and this person plays a lot of Metallica.

Speaker B

This this is sort of outside the scope of Musimap, but I'm I'm sure that both you Garen and Ty have a have an opinion on it, or, you know, can sort of comment on the future of this, but digital identity in general. Like, you know, we are inherently digital people. some probably more than than than others, but we are a digital generation. So how does that play into systems understanding us and and system being part of our lives in a different way? Do you think this is an essential part of it? Like what what's the big commentary? Can we tie it together with other digital identities? What's the future roadmap of this?

Speaker A

well I might say this. So, you know, obviously there's a professional identity. Let's talk about that. Like the musicians, the composers, the people who are l basically public figures that are part of recordings. This producer produced Quincy Johns produced this track. So that information is part of the credits of music. And generally speaking, it's not sacred. And and you have to be kind of okay with that being out there. So you're a public persona at that point. Now, how much beyond that? Like Do I know where you're from? Do I know where you live? Do I, you know, like that's all personal identificable mission. So you can't really have that. And and I do believe that these systems are really about using that information to help you understand what kind of music or media you like. I like films produced by Quentin Tarantino, you know, like and he's a public figure. So that's one thing. On the individual side, it's really has to be completely up to you. You have to opt in to share any information. You have to like, you know, what the rules are for dealing with people who are under the age of 18, what All that stuff is kind of being defined kind of on a outside music basis on a bigger global basis. But it's important because you know, this this is kind of up to each person and people feel different about differently about it. And so, you know, our system will work with whatever information we're allowed to have and and and we'll do the best job we can. And if we don't know, you know, anything about the kids in the Mercei, you know, then we won't be able to personalize anything for them. Unless you speak to the agentic voice in the car and tell it to do so.

Speaker B

Gorin, I think it's quite interesting because you have a background also as an investor and being part of strategic advisory and advisory of startups. Do you see a general innovation level close to what you're describing now? Is that very rare in music or do you see a lot of startups actually attacking things from novel and very impactful angles in general?

Speaker C

Unfortunately, there's a lot of things that I see that some very smart people come up with sort of solutions that are interesting but not that useful. So when I get involved, I like to make sure that the team has people involved that have deep understanding in what real problems are. So we can actually build something useful. So now with Muslim app. We have, I think, a fantastic team, everything from the musicologist, some great tech, but also great understanding of where there's real problems in the industry. So as I hinted earlier, we have some fantastic products like music profile, which you don't actually need to know who that person is. It could just be a code, right? And that's then something that we know is going to be extremely useful for music fans. Then we have by really being obsessive with the amount of data we collect. And now I'm talking about ISLCs, ISWCCs, who is playing on these records, who are the composers, etc. Then we can also do, you know, as you might remember there was this attempt to do the global repertoire database that sort of failed. Not for lack of trying, but maybe it was too much of a committee. And also the time was probably not right. It was probably too early. So we actually believe firmly that we can be an industrial layer of truth when it comes to solving these issues. We're not looking to be the clearinghouse. What I really like with music tech is Companies that enable. I'm not a big fan of disruption. So if if the mission would be to go in and you know put PROs out of business, I think that's not particularly meaningful. these are great organizations, actually membership organizations, but they might not be the best with the tech. So we can actually then really solve some of these real serious issues for the industry. But as an enabler, as a partner, a service provider. So this is when I look at a music tech outfit, that's sort of what I what I love to see: a deep understanding of what really truly is needed. The right players in the team to be able to actually execute on this. And and that's that's what really draw me. since day one to muse the map and I've been, you know, engaged with this firm a couple of different times and and now we are, you know, so excited to do this relaunch.

Speaker B

And that plays into a bigger question and maybe one of the final questions of the interview and thank you for for being so generous with your time, is will this change of the industry, this embedding of experience come from commercial actors alone? Or will they be a big part of sort of a collective movement tie what's your what's your take on this?

Speaker A

Well, I would really like to see music the the main product that people do, which are music subscription service, I'd love to see them evolve to have a better product. It's it's unfortunate for me that like the video platforms like TikTok and others, there's a lot more innovation occurring there, audio visual innovation occurring there. And so it's kind of divided the attention of the fans. The fans are over there looking at all the video, a lot of which has music versus in the main product, which actually is supporting the music industry in a lot, I think, better way. So there's I would like to see innovation there. I think that audiovisual innovation is important. I think that emotional understanding of the fans to get a better connection is important. It will, I think, become super important because what will happen is these other forms, the other competitive forms out there will continue to grow, which ca will cause the I'll call subscription music service industry to have to evolve because they have to compete. And so, you know, hopefully we're there with great technology and tools to really help them evolve. We have a lot of understanding ourselves, a lot of experience. So, you know, Music Maps in a great position to really contribute to this effort across a number of different areas. And so I'm excited about it. And and you know, I was there at the beginning of iTins. I was there at the beginning of Apple Music and Grace Note had a huge contribution to those platforms and products. So I've seen it before and I'm thinking I'm seeing it again. And I'm hoping that we can make this happen because, you know, music is so important to cult culture and society, it's too too important to to not improve.

Speaker B

So I think this will be the last question of the episode, but it's it's very much in the spirit of the podcast, which is a music technology podcast first and foremost. Ty, you you've you've built and sold a company before. you have have a sense of what things are worth. Do you believe this can become a major company? Can this be, you know, a unicorn? Can this be like very impactful for the industry long term? And the reason why I'm asking these questions is I think it's very important for us to understand as entrepreneurs. Where are the opportunities? Where are the where are the opportunities not? Which type of businesses could potentially make it not? What what's your commentary on that? If you want to look from let's say P classes or or V C classes or that sort of angle?

Speaker A

Yeah, I mean, so you know, I can just tell you over the I guess it was a Grace Note eighteen years, something like that. you know, when we started with Grace Note, we had the C D recognition database. That's what we started with. And when we when I left Grace Note and it was sold on to, you know, Nielsen, it was all the music data that you might imagine, but also controlled most of the information around movies and television, controlled the billboard charts, it had the had the Olympics. Sports, you know. So in other words, none of those things were imagined at the beginning the company, but but we are we realized they were all related. The data about all these things was related. And so when I see where Musimap is going, it's kind of like, you know, we're starting with music, but there's emotional value in video. There's emotional value in other verticals that we'll eventually get to. And so I feel like the it's unlimited where where it can go. I'm super excited about it. You know, I really wanted to have actually Echonest as part of GraceNet. We tried to do this back in the old days. And unfortunately we weren't in the spot where we could make it happen and they went into Spotify. But for me, this is like a second shot at getting to have this emotional and audio intelligence layer, music intelligence layer, which will become just an intelligence layer later. Won't just be about music.

Speaker B

Amazing. Duran, Ty, thank you so much for your time. it's gonna be so exciting to follow this. It's definitely a big new interest of mine understanding music data, consumption data. I think it's really interesting what AI is doing now with toolings to making that more available for non developers. and I think that's quite important for the music industry as well. Thank you for your contribution, both in your in your lifetime and and what you're doing now.

Speaker C

Thanks for having us.

Speaker A

Yeah, we really enjoyed it. It was great. Thanks. Mm-hmm.