Decoding AI's Role in Music

When you can text prompt a song into existence, then if you have some kind of ambition as an artist or a creator, then at some point in that chain, you're using a machine instead of where previously you'd have used a human. AI, AI. Right now it seems like it's the only thing the music industry is talking about. But it's really hard to navigate through the noise. The conversation we'll do today is to do exactly that, navigate through all the noise with AI.

Hey guys, and welcome to another episode of Sound Connections. Today I have a great guy in the studio. It's Drew Thurlow. Drew, you are very, very passionate about AI. That's something you know a lot about, and you have a very rich history in the music industry. Welcome. Thank you for having me. Yeah, I'm excited to get into it. So, Drew, just to set the scene of who you are and sort of what perspectives you bring to this conversation, could you just very briefly run through your career history and what you've done until this point? Yeah, I, it's been in music a long time.

I started my career in entertainment actually working for a major league baseball team, the Boston Red Sox, where I worked on booking anthem singers, doing some music initiatives, but I really wanted to play music. So I moved to New York, started playing in bands. This kind of coincided with the rise of the indie rock community in New York. So I had a lot of success, played in a bunch of bands, did session work, got signed to Atlantic. After a few years of that, I got really curious about the business creative side.

So I took a job at Warner Music Group at an imprint called None Such, which is a really kind of cool, eclectic imprint in Warner Music Group where I did marketing and A&R. Then around 2011 or so, kind of the low point of the recorded music industry. It was frankly a pretty depressing place to work at a label. The, you know, the keeping your job was the new raise. It, iTunes had...

Jakob Wredstrøm (01:58.006)You know matured and it was the only game in town, but it didn't nearly make up for the decline in physical media Sales and streaming hadn't had just launched in your part of the world not in the US yet actually that came in 2010 11, but So I went to work for Pandora and when I got to Pandora I realized I was the first music hire at the company in 2011 and I had an extraordinary five six years there Doing everything from talent buying to content licensing copyright law lobbying and US Congress. And it was an extraordinary kind of six months, six years rather. And then as the company matured, well, sure enough,

thanks to streaming, the record labels were healthy again, doing better. So Sony had recruited me to help relaunch their masterworks division, which historically had been a kind of a classical jazz imprint, but in the era where the pop labels were doing basically just hip hop and kind of mainstream pop. You know, they didn't have anyone at the music group doing different kinds of music like Americana, rock bands, electronic music that wasn't Calvin Harris. So I was head of A&R for Sony Masterworks.

I did spend five or six years there. Left this past January to see what else was out there. And it turns out all anyone wants to talk about is AI, or at least mostly. So I spent a lot of time doing some other things. I have some other consulting clients. I consult with the NHL on their music and entertainment content. And I have some more traditional distributor clients. But I spent a lot of time talking about AI, a lot of writing about AI. And it's been an unbelievably interesting year. It's a lot to follow. And it's a lot to keep on top of. But we'll try to get through as much as we can today. Yeah. I very purposely have not talked about AI in this podcast that much.

And that is because you need a certain perspective and a certain mindset to actually have a proactive conversation that actually leads anywhere. Because AI is so complex and it's so quickly moving that whatever you believe is hot right now is gonna change. So the only way, in my opinion, to really talk about it is the broader concepts, the broader movements, and sort of almost the philosophies behind it and what is happening and how we need to think about it as an industry, more than specific small cases of a hype. You can sort of...

really sort of not compare it to what happened with NFTs a couple years ago, but that was almost like a precursor of some of the hype that's happened with AI. And that's sort of why I see around it. But I think today is quite exciting. I've been reading through your LinkedIn posts and content, and I really love the take that you bring to the table. So I'm very much looking forward to the content we're gonna talk about today. But if you were to bring like an overall commentary on what's happening with...

in AI, especially in the music space right now, what would that be? Yeah, there's been a lot of chatter and reporting on AI, and you're right, it's kind of overwhelming. A lot of people are really freaked out. I'm here to tell you, don't be freaked out. There's been also a lot of comparisons to the blockchain, NFT, crypto hype, and then subsequent crash. I'm also here to tell you, AI is not NFTs. It is not Web 3. It is so much more turbocharged than all of those.

There are some kind of historical examples in the music industry we can point to. I think sampling and the rise of hip hop culture in around 1980 or so is actually a pretty good comparison, at least on the copyright legal side. But there's a lot of different facets to AI and I think we can talk about them in two ways. The way it's affecting creators and musicians and then the way it's affecting the industry and the business side of it. And then relatedly copyright and that.

So I think we should start with how it's affecting artists. And I'm here also to tell you that AI has been around for a while. Machine learning has been around for a while. And I've spent a lot of time in pop recording sessions. It's a technological wonder what happens in studios these days. The rise of digital recording technology, which started pretty earnestly around 2001, 2003-ish with the rise of Pro Tools and DigiDesign. A lot of pop recording sessions are utilizing AI and machine learning in the creative process already with plugins and soft synths. Most recordings you

hear these days commercially released are using some sort of AI with mastering and mixing. There's a product called Ozone, which is an iZotope Native Instruments product, very, very heavily used now in music, has been for years. It's AI. When you. a lot of these drum programs that people are using to sequence beats and use different kinds of mathematical patterns to create beats. That's an example of machine learning and AI. So it's been with us. And I just think understanding that is helpful and comforting and then extrapolating the advances in the tech about what's next, I think is also important because even though it's been with us, it is a little different and it's about to be way more pervasive. That's great. I think.

One of the things I really experience a lot is fear. And I was at a conference recently last month where 40 music industry professionals with a lot of experience, I was lucky to be there. Really talked about AI and the prevailing feeling was fear of the unknown. But also there was a sense of, this is probably for the good. But where do you think the fear stems from? I'll give you a personal example. I am a session musician. I spent a lot of time learning how to play my instrument.

and I know theory, and I was primarily an electric bass player, but I also did a lot of session work on guitar and, you know, more limitedly on piano. Some of these generative AI tools make it so I'm not needed anymore. And as someone who spent a lot of time and years learning my instrument, I don't know if that's awesome or if that's terrifying, but you can make really competent music with these machine learning and AI tools, and it's directly threatening.

artists and their livelihoods. I'll give you an example. I used to make a lot of work getting paid to play bass on sessions. I'd get called in to play, and I would get either a union rate or I'd get some kind of money, learn a song, bang it out. Sometimes it'd only be a couple hours. You can now replace my bass track with an AI-generated bass stem, and then you can use AI to edit it to do all different kinds of takes. So that's a one-to-one example as a human is no longer needed in the recording process. And then you can extrapolate that to the entire recording process, entire songs. When you can text prompt a song into existence,

then if you have some kind of ambition as an artist or a creator, then at some point in that chain, you're using a machine instead of where previously you would have used a human. Now, there's examples of history and technological events all over the world where this is just the case and this is just another example of the adaptability that's needed. If someone's making their livelihood playing instrument or scoring a indie video game or writing the library music for TikTok videos, then AI is here as a direct example of taking away your livelihood. So I think that's where the fear part comes into play. And frankly, it's justified. You talk about the historical events that sort of have

been similar and comparable. Could you dive deep into some of those? Yeah, I mean, outside of music, like I think everyone's having the same argument around the discussion on self-driving cars. It's a very, you know, driving an automobile or a truck is the number one employed career in every US state. And you know, we see what's coming with self-driving cars. So that's it. Not only is that personal livelihoods disrupted, but incredible economic disruptions are coming.

And in a music context, I mean, there's been plenty. I mean, there are people who, there are a lot of people who lost their jobs when the CD was invented because vinyl pressing plants went out of business and now we're seeing them come back. So maybe everything's cyclical. I was working at Warner Music Group when Tower Records went bankrupt and closed and thousands of people in the United States lost their jobs working in music retail.

and all the infrastructure that went around with labels being companies that manufactured and distributed products rather than the media companies they are today. So there's certainly a lot of examples of technological change happening and wiping out what were previously durable industries, but that's just kind of the nature of technology, music or not. Yeah. I think a lot of the fear also lies at least from the industry, sort of taking the damage it did to the industry, I know, 20 years ago.

if that's going to repeat itself. What is your predictions or assumptions about the music industry economy at large and the effects that has, how it's affected by AI? Yeah, I mean, the music industry's had a pretty decent six years, but before that, it was a pretty awful 15. 80% of the global workforce who worked at record labels lost their jobs, and I think people are definitely ready for the good times to continue and don't want to revisit that era. At the same time, there's like some very illustrative examples of why that happened the first time. And I don't think they're going to be repeated the second time. I mean, when, when Napster proliferated,

the major labels reactions was to try to sue them to oblivion. And people don't remember this, but very late in the Napster era, Sony music and Universal Music band together to create their own kind of file sharing paid platform called Press Play. And it was just like way too late. Naps, not only had Naps are come and gone, but there was also like, you know, the 15 or so comparable file sharing services like Kazaa and Bayer or something, or what I don't even remember. Then all reached scale. It was way too late. Major labels have done a really good job. And, you know, I shouldn't leave out Merlin and the Indie Label Consortium too. They've done a really good job of being smarter about technology this go around.

And now major labels employ what they're called partner management teams who, when they find emerging tech or they find some kind of new platform, then they work with these emerging techs to kind of develop their businesses so they can return really good customers and a new revenue stream to rights holders rather than trying to sue them to oblivion to protect what they have.

So I think the major labels are really smart. And you see that with some of the partnerships that the labels and rights holders have done with early AI companies now. They're smart and they're really good and they're kind of testing the waters. So the first major one was a company called Endel, which makes kind of like sleepy meditative versions of contemporary music. And it's a startup that had been around for a long time, but Universal did a deal with them and now they're working with the other majors.

It's smart because you're basically you're using it to create new IP and this new IP is not competing with your what the front line labels are doing like a sleepy Song that you listen to while you're meditating is not going to compete with like the radio banger that you're trying to you're trying to release So it's smart and then Warner music group actually was Sony who just did a deal with a company to remake and remaster and remix some Pink Floyd material that fans can interact with this Pink Floyd material in interesting ways. They can remix it, they can play with the stems, they can create new video content around it. It all kind of lives in this self-contained app or platform. And it's not really utilizing Pink Floyd music that was

particularly popular. So it's not really cutting into the real revenue that the labels are making, but it's giving fans something they want, which is to interact with the music in a deeper way. So these are kind of baby steps that people are taking on the industry side. And I think there's so much more to come with that with regards to using AI to either create new IP or create consumer tools, but they're being smart about it. Then the labels know that the opportunity that exists is matched by the possibility of disruption and they really wanna be on the right side of history this time. That makes a lot of sense. And that's also sort of the general narrative that I hear out there is people really wanna play along this time.

How's this going to affect us? Can you give us some use cases that things are going to change? You mentioned you being a session bass player, you might be sort of replaced in some scenarios. Can you give me some more scenarios of what will probably be replaced through AI? I think one of the things that I spent a lot of time thinking about is the music industry and the current infrastructure of the music industry was designed that there would be a certain amount of music in existence.

And when I was at Pandora, we had the history of Western recorded music at our disposal. It was about 12 to 14 million songs, if I remember correctly. And then the challenge of Pandora was how do you serve the right song of that 12 to 14 million to the right consumer at the right time and the right mood? Well, 12 to 14 million songs are now being released every three months.

I mean, no one even really knows the real number. It's some shockingly huge number. There is now over 200 million songs in the kind of like, RIAA database of commercially released music. And that comes with a few challenges. Number one, the distributors and the DSPs cannot afford to host all that music. It's too much. Number two, consumers don't wanna hear it. I recently asked some of the followers on LinkedIn they could guess of the 200 and...

plus million songs that have been commercially released, how many of those 200 million actually get listened to more than a handful of times a year? Do you want to take a guess? I'll say 20, 20 million? 600,000. 0.3%. Wow. Less than 1% of the music commercially released is actually wanted by consumers and actually gets streamed enough to like pay for the hosting costs. So there's just too much music out there.

And people, so now the DSPs and distributors have a real question to ask, like, how do we manage this? So, and that's only gonna be gonna explode with the way these generative AI tools are making it easier and easier for people to create music. So I think one of the kind of insights that I have is up until now, the valuable piece of music has been the song. It's been the output. So the writing and the recording process has been a sunk cost to you can get to the song.

and then you can monetize the song in whatever format that takes. Well, in this future, there's gonna be hundreds and hundreds of million people making music on these gamified kind of fun gen AI tools. It's pretty clear that the distributors aren't gonna take it, DSPs aren't gonna carry it, and the consumers don't want it. So then the opportunity for the music industry is in monetizing those creators. And I think about it like the way, think about it as like an Instagramming of the music industry. Like none of, I mean, very few people like bought I don't know what aperture really is, but billions of us loved our Instagram feed. I think there's about to be a whole bunch of people making music for the fun of it and tweaking it in ways that

are interesting and doing it in a gamified way, as I said. The opportunity is to monetize those people and not necessarily what they make. There's a few platforms already doing that at scale. BandLab, very big, interesting company in Singapore. They have over 60 million users. They're serving kind of a prosumer audience. I don't really know, but my bias is that BandLab is serving a customer that actually kind of likes music and maybe even plays an instrument. There's another platform called Boomi, which is the complete opposite.

You basically text prompt music into existence. You can name it and then you can send it to the DSPs. So they've made 20 million songs that way that are now sitting on DSP servers. How do you encourage that behavior? I mean, I think people making stuff is good. People making art is good. But how do you do it in a way that's responsible for the music industry and how you do it in a way that's serving that consumer class? That's interesting. Can you walk me through the revenue here? So you're sort of saying we can monetize the creation of music rather than the consumption. Sort of I'm paraphrasing now, but that's sort of the essence of what you said.

The journey of that money, the journey of that monetization, who does it go to and how does the flow go? I think that's what people have to figure out, right? Like, think about all the ways that you would make music and think about all the ways you would do it if it was like a fun, free app. So, certain things you could do. You could sell, you could do in-app purchases around like different sounds and beats.

You could probably license in a well-known musician's AI vocals and pay for that. You could probably charge for some kind of hosting or posting fees. You could probably upsell people to a video platform where their music would be synced to video, and then you would sync those two together and you could put it on your socials. There's all different kinds of ways that... You think about it less in a music production capacity and more in a gaming...

upsells in an app capacity. Like, I don't know, I've never played Candy Crush, but like, I'm sure there's like upsells on Candy Crush. I don't know what they are. Like maybe you can like change, change dress your avatar or change the color or something. Like there's all different kinds of ways. Maybe you, maybe the app, maybe you pay a subscription fee and you get 32 tracks instead of the basic 16. Maybe I'm giving away some like really good startup ideas here, but like.

I think the point is that consumers are going to play with this stuff. The technology is here, it's happening. So responding to the market is going to be the opportunity. So that gamification of music, there would be some underlying rights holders that probably would get paid through some sort of licensing we haven't figured out yet. And then obviously there needs to be some monetization of the platforms itself. But if you can break it down even more, again, now I'm just doing a hypothetical example, but let's do a thought experiment.

So we have this platform that has this generative AI aspect that sort of enables people to make music for would be the people earning money on it besides the platform itself. So how are they making music? Is it text prompt based? Is it kind of just like algorithmically based? Yeah. Okay.

Yeah. Let's do that. Well, there's a couple of options here. Like in the boomy scenario. Bumi owns the IP and the person licensed the Bumi licenses the rights for the person to distribute it and then they split the streaming revenue. Now virtually no one wants to hear this music but when you release 20 million songs you and your mom and your friends are going to play it once so if you can get 10 streams on 20 million songs then you're that's 60 million streams and maybe you can build a business around it. I don't know. So there's that example. There's like

The upsell within the process example that I just kind of articulated. There's also like other ways, maybe this music ends up in a sync library and you can sync it. So there's all different kind of ways that you can interact with the music in a way and then throughout that chain, you can monetize it. It's very much, you know, this boom example is the platform, is the company itself that earns money.

That does not necessarily the existing industry and the existing artistry and IP rights sort of could be receiving any benefits from that success. You're right. And I think we should make a distinction. I do not think that the next 20 years of pop stars are going to be avatars. Like people's relationships to their musicians they love goes so much deeper than the music they make. Like I don't think people want like an anthropomorphic Taylor Swift. I think they want the real thing. So we're talking about.

In a lot of ways, we're talking about a completely separate class of creators and in turn, a different revenue stream for artists, for the industry rather. So yeah, let's make that distinction. I do not think, nor does anyone I know advocate for replacing humans with bots. Yeah, that's an interesting talk because I've emerged to agree and I saw some talks about this. Can you just work through the, maybe let's call it the psychology of fandom?

When it relates to AI and why you're not sceptic for that taking over. This is something that I've thought about recently, and I just read a really, really interesting kind of treatise on this, and I didn't know this, but when the Duraga type and then photography was invented, that was actually quite controversial. That was a point in history where a lot of artists were making their living doing portraits, and all of a sudden you have this technology comes along and you can replace the portrait artist.

And that freaked a lot of people out, put a lot of people out of business. But really what it led to is it led to impressionism and it led to those artists being freed up to try different kinds of techniques. And you have like the whole, you know, the history of like 19th century and 20th century art is now is like really a reaction to portraits, not being the dominant form of art artistry anymore. And I think that's a really interesting insight into creativity and the human condition when it comes to creativity.

So I'm really excited to see where does like, if you can make the basic kind of music really easily, where does it push creators and how do they use these tools to augment that? So I think that's really, I'm really excited personally for like that development. And I think context really matters.

And you think about any kind of piece of art that humans love, whether it's painting or a film or a TV show or something, the context matters. Who makes it matters? What the subject matter matters. Making something without context that's devoid of any feeling or emotion has its place. I don't think it's the kind of art that people are going to want to consume. One argument against that, and I'm not saying it's valid, but one argument you could make against that is the whole playlistable consumption of music where you're a passive listener in the background.

And I guess I don't have any statistics on it, but you know, on a pretty avid chooser, you know, a lot of their consumption would be that. Where they don't necessarily have a relationship with the artist, you know, just driving the way over here to the studio. I was listening to some music based on a Jeremy Soccer playlist. And I was in my car and obviously I didn't really look at who the songs were. And there was some songs I liked and some songs I didn't like, but I didn't really have any context or relationship with it. Do you think in that example, What an example. AI would go so for that. Yeah, I do. I think that's a really good example. I think that the anonymizing of artists based on playlist culture is a real thing. And I think there's some

Durability to real artists, but yeah, I absolutely think you're right. I think that is a... And I'm even more bullish on AI being able to make really interesting, credible music more than most people. And one of the examples I tell is that when I started researching this stuff in earnest in February, March, you couldn't really make an acoustic guitar sound credible. And as someone who's kind of deep in the songwriter community, I'm on the board of the Americana music association and I spend a lot of time in that community. I think that was a source of comfort, but by May, that wasn't true anymore. You can make a pretty good acoustic guitar sound now. And, you know, and the argument of like, but the lyrics are human condition. Like.

I don't know, man. Chat GPT can come up some pretty clever, pretty interesting thing that feels pretty real. So while I'm bullish on the personality is still mattering, I'm definitely bullish on the, you know, I don't think humans are so genius that a really sophisticated. AI system, maybe that's based around quantum computing can make something that we all find really compelling. You're right.

One of my, let's call it opinions and commentaries about exactly that point is sort of, the music industry at large is going into a new hot topic, which is fandom. And obviously this has been very hot in the East for many, many years, but going more and more here in the Western world. And if you look into examples that I've worked with is, maybe that's called dance music, which is a great example of artists that might have millions and millions of plays, but they play a show and no one shows up. Because it's sort of based around playlist culture.

And those artists in the future would really be threatened by AI. But artists that really focus on creating community around them and really resonates with a certain niche that has a very personal relationship with music are probably the ones that won't be as threatened. Do you think that's correct or do you want to... Yeah, no, that's a good point. And I think that brings up some really interesting illustrative things about how that's evolved.

The streaming era, ticket sales for live and music consumption was super correlated. It was like, if you sold this many records or you sold this many iTunes downloads, you sold this many concert tickets. And it was like pretty well correlated. That is completely out of whack now. There's basically not correlated at all. You have your superstars that can do both, but you have lists and lists of people who stream like crazy who can't sell more than a couple hundred tickets in major markets. And then you have people that...

Jakob Wredstrøm (28:54.806)can sell really well live, but no one's consuming their music on streaming. And, you know, previously that was only like jam bands that were in that category. So, um, so yeah, it's a really interesting time about like what it, like how people consume music. And one of the, one of the disappointments of streaming is for the last. You know, we were now, we're now, you know, 14 years into streaming, launching and at scale and maybe 12 and then the whole world.

And for years, if you weren't a top charting artist, you were told like, just be patient, streaming is gonna grow, it's gonna mature, it'll be there for you. You'll find fans, you'll be able to monetize your career, especially in genres that weren't like mainstream pop. And here we are, streaming is really kind of plateauing and it has not returned the value to the artists in large parts of the industry. So that's where you mentioned the kind of the whole fan monetization thing.

Like super fans have not been monetized properly. A lot of people talk about that. There's a lot of startups trying to work on that. And I definitely think there's opportunity there. And that's another example of maybe AI helping. There's a lot of companies that I'm aware of, they're starting AI data analytics tools to reach these fans so you can communicate them directly, so you can come up with different ways to engage with them and monetize them better than just taking their 999 streaming. I will say too, There's this whole push in the streaming ecosystem about artist-centric royalties. And like, they're just, they're not like even, even the changes that were just made to from Deezer

And the Universal Music Group, and then with like not paying out songs that get under a thousand streams a year, like they're not artist-centric. Articentric is when I, as a consumer pay my nine 10 99 a month. And if I only listen to that one artist that month, they get all of my subscription. That's artist-centric. That is not the model, even though they're calling it that. So I don't see streaming changing the calculus at all. And I think the path for some real creators to be musicians and make a living doing it is in the monetization of your fan base in different ways for different tiers. Yeah, and it's sort of returning to the old way of building a career, right? And I think that's kind of interesting.

Moving forward as an artist, I think being likable is super important and having personality and sort of standing out from the crowd and you can't just build a career around the music itself. There needs to be a concept. I actually- There needs to be a personality. You're absolutely right. And I spent a lot of time talking about this too. Just cause you make something doesn't mean you get to call yourself an artist. And I've reminded of this when I was playing music myself, when the blogosphere was all the rage.

And you could make some kind of like bedroom recording. And if Pitchfork liked it or Stereogum or Brooklyn Vegan wrote about it, well, like you could have 500 people come to your next show. Which is insane to me as someone who started playing in garages in the 90s. And I had been playing in bands for a year before I actually got on a stage in Providence, Rhode Island where I grew up.

And I think I had been playing music for 10 years before I set a label deal and probably 15 or 12 before I actually did like a national TV thing or played bodhr or whatever. And he was in, in the early 2000s, I remember very distinctly remember Beirut who's a kind of any rock, any musician in Brooklyn and did, did had a good run in that time and is still making music. I remember he was like third or fourth show or whatever.

He made a really cool recording and then his third or fourth show was at the knitting factory and there's 500 people there including like David Byrne and David Bowie in the audience. He was like scared to death. He's like, I've been doing this for a few months. I haven't even figured out if I want to do this for a living. Never mind like getting my stage show together and figuring out how to be a performer. Like it's just, it's a lot. And you know, he stuck with it and he's made some cool music with like, with lots of kind of disappearing moments in between. But I just...

As technology comes down and it gets so easy to record music these days, it's like you don't deserve to be a career artist just because you make something. Like you have to work at it. It's a sacrifice. People don't like take their first job out of college and finance and expect to be like a CFO the next year. Like the same thing comes with artists. Like you have to work at it. You have to make something compelling enough that people care.

Your visuals and the way you interact with the public has to match that, you have to be good at it. You have to stick with it. And then if you do all those things, maybe in several years you'll have an audience that can support you in a way where that's all you're doing. But I really have no patience for the people who kind of make something quickly on GarageBand and release it and then kind of have expectations around an audience or a career just because they like it.

Or they want it. Those are the people who are usually in it for the wrong reasons. The audience for this podcast is sort of intended to be people who are curious about entrepreneurship and music industry. And I've lately thought about founders and artists sort of the same way and try to use sort of analogies between them to understand each other. And one of the things we briefly talked about, I run a venture studio where we're like a founder-centric company, which is about, you know, pairing ideas.

With founders and everything in between. And one of the things that I did wrong in the first part of the journey was sort of assuming that people who were skilled and willing could be founders. And sort of, and it's true to a big extent, what I realized is sort of founders also needs to have sort of the X factor like an artist, right? You know, they need to have something special to bring to the table. They need to be extremely passionate, extremely knowledgeable and sort of that uniqueness in them is a part of the journey. But then vice versa, I've also made the mistake the other way, right? Where it's like, well, if you're a skillet as an artist, or you have some sort of charisma, it's enough for you to pursue that journey. But you need to also look at the star founders, right? The star entrepreneurs,

What do they have? And sort of those parallels are also evident. And for me, that makes sense, and hopefully also for the listeners. But I think what it really comes down to is like, do you actually have something incredibly interesting to bring to the table? And if you don't, maybe it's a very serious hobby on both the founder side and asset side. I think we as people want to be moved and we do not get just by the ordinary. Maybe you can tell an ordinary story in an extraordinary way, but there needs to be something very, very special. And I think maybe that's sort of what AI is forcing us to realize faster.

Interesting. Yeah. I would agree with you. I think there's two things that- artists and founders need, because an artist is really an entrepreneur themselves. All a label is a VC firm for musicians. So I think the two things that you need are, number one, you need to have a unique view, special to what you just described. Number two, you have to not quit. It's amazing to me how the best lesson over and over again for entrepreneurs and for artists is just like, don't stop. It's crazy. Some of these startups that do really well.

You realize they've been doing it for eight years and they just they don't quit. And the same thing with artists. I've artists, I just, some of the artists I remember just like playing for three people, the ones who didn't stop after 20 years and doing really well. Like that's just, that's, I can't stress that enough, but it's interesting. Yeah. When, when you can, when AI can come along and grab the low hanging fruit and make the easy stuff, what it make it go from easy to free or seamless, then what, then what, what are you left with? Yep.

Then you need something really, you need to really be providing value. And that's one of the things I talk about. I do, I have a lot of music tech startup clients and I'm really kind of researching this space a lot for some upcoming writing projects and you really do need to be providing value to the ecosystem. You need to have cutting edge tech. You need to really like be solving a problem. And if you're not. Then.

You just, you don't have a future. And it's okay to make a mistake. And it's okay not to do that at first. But then again, don't quit. There's a lot of music startups now who still have funding who were like web three NFT companies and that crash and now they're all pivoting. And it's been really interesting to see which ones are pivoting and which ones are throwing in the towel and which ones are like really hustling to make it work. And which ones are kind of like, ah, I still got some runway. Let me see what will stick.

I think this is a really interesting conversation and I haven't really discussed this narrative that much when it comes to, you know, we need to be exceptional as creatives and as founders. And if we're exceptional in what we do and what story we tell, there's probably more longevity to it than we would think. Because at the end of the day, you know, our society and our reality consists of humans and we want to relate and resonate with humans.

Um, and I think AI is a part of the equation, but, but I have a hard time seeing it be a real competitor to real interactions and relationships. I mean, like, this is why I was so struck by the photography example. Okay. Like what's the musical equivalent of a portrait of, is it like library music and short form video music for a TikTok? Okay. Well, like what else is out there?

What else can you write? Like the higher level stuff cannot be replaced with AI. It can be assisted. There's a lot of kind of tedious annoying things you have to do in the recording process, like stem labeling and premixing and like trying out sounds. And then there's a lot of interesting ways that AI can help reach an audience or write a marketing plan or book you, send out pitch emails to...

500 journalists, but you're right, the actual core of it is something that's going to retain its kind of value as a human entity. Yeah. And there's so many aspects to AI that I'm so positive towards. I used to be a songwriter producer myself and sort of my creative aspect of, or approach of creating anything, if that's a business idea or a partnership or pitch or whatever. For me, my limitation has always been speed. Because I'm...

I can think fast and I can have a lot of ideas coming very quickly after each other, but with sort of taking those ideas and taking that creativity and putting it into something concrete has always been the challenge. That's something I can use AI tools right now to do to really increase my speed. And I also think that's, you know, many creatives work like that because they have an...

Overflow of material and impulses that they need to bring out to the world. And through AI technology, they can sort of put that faster into the world, which also means that they don't need to stop all the time. And somehow sort of remain that creative space and be effective. Yeah. That's another thing about being a creative these days is like, you can't go away. You have to continue to keep what I call keeping the faucet on. It's the advice I gave my artists all the time. Like I'm sorry. I know you hate social, but like you have to keep at this. You have to keep writing and recording music.

Our release cadence has to be every month. And it's also this idea of being creative means you're unpredictable and you're messy and you're disorganized and you're a genius waiting to be struck with it. That's not real. And I'll tell you an example. My all-time favorite artist growing up was Bjork. And then I worked on two records with her. And...

I will say she is all as amazing as I thought she was, both creatively and just a really great person to work with. But she's one of the people that I've considered probably if up there in the most creative, innovative person in pop music the last 20, 30 years. Working with her closely, she's a CEO. She's very disciplined. She's very organized. She had ideas, but was very tenacious about following them through.

Is an underrated and underreported part of creativity, is like collating and managing and focusing your ideas in a way that's like digestible, releasable, and has commercial value. AI right now, could you just give me a status on what's happening in the AI space within the music industry right now, and also what you believe what will happen within the next 12 to 24 months? Yeah, I definitely think that people overestimate.

The effects this will have in the short term and underestimate the effects long term. So next 12 to 18 months, what's going to happen? Not much. I really don't think much. I think you'll start to see around the margins some certain things, but the music industry and the ecosystem as we know it will continue. Long term it could look quite different.

Let me take you through an example. So I think there's bound to be some kind of... pop star who's never ever picked up an instrument or sung a note in their life. And some of the pop stars that we know and love today, they're not the most creative and they don't write their music. But we still have the, we still have the lineage of like the boy band era where like you were at least like a child performer and you were a singer and you performed on your own and you really, you really seeked out.

The kind of like lifestyle and, you know, be, being musical was part of that. But I think we're very quickly in the next few years are going to have a pop star who like picked up one of these. These gamified apps. And because of the way that they understand music or think about music or the, what they've been influenced by, they're going to be able to make some really kind of cool music and they're like, they're never ever, they've never done it before. And I think it's going to be, people are going to find that really compelling. And I think that's okay.

I think a record company is going to invest in that person. And I think they're going to go out and do a tour and a stage show that's heavily related to technology, including like vocal sampling and including like all backing tracks and session musicians on stage. And I think the stage show is going to be amazing because as we've seen with the sphere in Vegas, the production value is exploding and like, you don't have to even listen, you could go to those U2 shows with noise canceling headphones on and still be amazed. So I think that's the modern pop star. And I think the soon to be modern pop star. And because learning how to play an instrument or write a song isn't gonna be a necessity anymore. That doesn't mean the people who are quite innovative in that

Area and still do it the old way aren't gonna have huge careers, that's not going away. But I think you're gonna start to see the definition of what it means to be a quote unquote artist expand. As far as the industry goes, I think you're going to see a whole bunch of dead artists making new music. And it's going to be really valuable for the rice holders.

And I know that the catalog divisions of the majors, which at Sony is called Sony Legacy, at Universal is called Universal Music Enterprises, and at Warner's is called Rhino, you're going to start to see Elvis training their entire Elvis. repertoire on an AI algorithm and outputting new Elvis music. That's going to be sanctioned by the estates and released like any other release. And we saw this just last week with Kiss, who played their final show at Madison Square Garden and then announced that they were becoming avatars and touring the metaverse in perpetuity. That was

Kiss. So I think you're going to be able to dial up. If you want to see Prince. I think in some kind of time horizon, which I will not embarrass myself and try to guess when, but I think you're going to be able to dial up a great Prince show that includes new Prince repertoire in some kind of coming time. And then the other really kind of insightful thing I think is going to happen is that you're going to get personalized music. And at Pandora, we kind of personalized radio.

It was, you know, it was, we would call the unicast. It was one feed completely tailored to musical tastes. Well, with AI, we can crack open the song and you're going to be able to create your own listening channel of personalized music and it's going to have all the musical characteristics that you love. And then you're going to be able to kind of give it feedback. Like you, you could thumb up and thumb down Pandora playlist feedback. And it's going to be.

Sometimes you're going to want this when you're going for a run, or sometimes you're going to want this for a dinner party, or sometimes you're just going to want it listening to your pop music. But I think we're very, very close to personalized music as a feed. And I think that's going to be really, really interesting. That is very interesting. But obviously there's a few steps before that. And this might be a boring question, but it's quite relevant. Right now, a lot of generative AIs.

Trained on content that's not being cleared. And sort of the underlying license that's been trained on need to somehow be compensated or regulated. In order to get to that almost utopian scenario that he just described, regulatory wise, compensation wise for the underlying training data, what's going to happen with those and how are we Yeah, that's probably the one thing we haven't touched on really in depth in this talk is the kind of government response, legislative response, regulatory response. I think most people assume

That if you take protected IP and you train it on a model and you create something new, I think most people assume that is illegal without compensating the initial rights holders. It is not. Right now, that is a very legal thing to do. This is so new that there's no case law that addresses it and judges can only rule on precedent. Part of it is that it's not clear that might just be considered fair use. It might be. I don't know that the industry, I think most people in the industry would say this probably should not be considered fair use, but right now it is.

There was an interesting court case in the visual art world that was an early ruling, which is going to continue to go up the courts and probably end up in the Supreme Court. But a... Some painter has sued Stable Diffusion and Mid Journey on your training an art algorithm, a visual algorithm on our art. And the output that you make, we should be compensated. Well, what the law says right now is if that output doesn't actually closely resemble any of the inputs, then it's not violation of anything, copyright or rights of publicity, which are kind of two different legal postulates that govern this stuff.

And it's true. When you have a million trillion pieces of training data, the output's not going to resemble any of the one trillion pieces of input. So how do you claim it's copyright law violation doesn't actually resemble anything that's copyrighted? That's a huge challenge. And we might need a whole new legal framework to address it. And the other part that's really challenging is how do you quantify musical inspiration? And if you take the Beatles catalog every song they've written and recorded and you train an algorithm and you spit out the you spit something out you just made Oasis like that's not illegal it's not even necessarily controversial but we have to we have to decide ethically what's going

To be okay and legally what's going to be allowed and once we have all that stuff I think people will start to see real investment in this space as crazy as AI is exploding now There's a lot of VC money sitting on the sidelines because no one knows where the law is going to land. No one knows it's going to be allowed. So if this is a Supreme Court decision that says, all right, sorry, and universal, but X startup can train an algorithm on your content because it's considered fair use, I think you're going to see a lot more VCs and the investment community jump into this space. And conversely, if the Supreme Court rules actually, you know what, you can do that, but you've got to pay royalty back to the f***ing

The initial inputs and you have to track what you're trusting your training data on and the training data has to be and the algorithm has to be transparent. Then at least that'll give us some business certainty, which is what people really want. And then the startup investment community can go follow the market forces from there. But the legal challenges and the legal uncertainty has been a really interesting and kind of understated part of all this.

Yeah, and so you add to the complexity. I'm no copyright expert, but there's also a difference between the basic ways that law in American, England works, contra the rest of Europe, where a lot of your laws are based on precedent and based on corporate protection, where a lot of it in Europe is individual ideal rights. And there's a difference there as well that adds to the complexities of that whole talk.

But what I think is really interesting is the underlying discussion whether or not that is actually not how music generally works. When you make a new band, it stands on the shoulders on previous material, and really navigating through what is actually top is super, super difficult. Yeah, it's a discussion I'm not very well-versed in now, but I think the whole industry is trying to figure out how to do that discussion. Yeah.

Something big in the tour buddies, I definitely try to do that. And it's no surprise that the EU is further ahead with not only their AI Bill of Rights and AI policy, but also even just last week publicizing new rules around streaming and how they can better serve the artists and specifically publishers. So yeah, everything is- Quite controversial.

I know. I mean, it was the EU and it was like, it's a legislative- It's a, they have a legislative function, but the, the actual statement and the actual ideas came outside of their legislative process. So I don't think anyone thinks it's going to got teeth to it, but it shows that they're thinking about it. And that's an interesting insight about how the EU is really kind of relies more on kind of personal freedoms rather than corporate protections, which is absolutely true about the UK and U S laws.

So you can extrapolating that, you can see like where the stakeholders want things to land and where the lobbying money is going. Interesting. But you know, as far as I understand you and this conversation, you're positive towards the impacts AI will have in the music industry. Yeah, I mean, yes, but also like it's here. Like it's not going anywhere. Yeah. This tech is evolving. It's been developed and it's having an impact. So.

There's always winners and losers when there's some kind of like sea change in any kind of industry. I don't want to be on the wrong side of that. And I also find this stuff pretty interesting. And I think that the industry is by and large is being pretty responsible about how they proceed because they've made a lot of gains in the last 12 years. And I don't think...

Anyone in the industry is in a hurry to have another, you know, dark winter. Yeah. And, you know, and also we can't forget, we're just, oh, long for the ride. You know, we're not the only industry that exists and this is going to disrupt pretty much everything in all industries and all aspects of life. Yeah. So, you know, us trying to control whatever's happening maybe doesn't make a lot of sense, but I really think making sure we're not left behind.

Should be one of our primary goals. Yeah, it's interesting. I do a lot of work in the synthetic speech vocal detection sector and the music industry is such small potatoes because in the synthetic speech world, like, okay, I'm like, I'm like copying Beyonce's vocals and I put it on YouTube and like it's legal. But like, Beyonce doesn't love it, but like, it's also kind of If that happens, you've gotten your bank account stolen. And if that happens, like a deep fake in your presidential candidate, you just lost a national

Election. So the stakes in the music industry are kind of comically low compared to finance and fintech and others. So it's also been interesting to try to follow the other industries and where they're at because they're freaking out a whole bunch more. I think that's a really important thing to talk about is, let's just calm down for a second. You know, we're not going to be the ones hit the hardest, probably not. And some of you, someone was willed, but you know, he's already speaking that happens all the time. I'm quite positive towards the change. And I definitely think that if you do not yet have a...

Opportunistic mindset about AI in the music industry, you probably should get it to some degree because there's definitely gonna be some winners and losers, but I also really think you can decide where you position yourself on that scale. That's well said, yeah. I write a lot about this stuff. I try to keep on top of it, so. Listeners, you should follow me on LinkedIn. It's where I do most of my chatter. Absolutely. And I...

You know, things come up and I address it. And I have a bunch of writing projects coming up that I could talk about in the new year, where there'll be more resources for this stuff. That's great, Drew. I really love this conversation. I think we touched upon some things that are not discussed as much as other things. And I think it's a really important narrative we've presented today. And I really do support people following your LinkedIn and checking out your content using your consultancy services.

Well, Drew, it's been a pleasure. I definitely think I want to have you back on the podcast, let's say in six months, sort of to comment whatever is happening there, because there's going to happen a lot of things and you really seem to be well versed and hopefully prophetic about some of the things that's going to happen. Yeah, thanks, Jacob. I enjoyed the conversation. I'd love to come back. Great. Thank you, Drew. Talk to you soon. Bye.