The Music Business Buddy
A podcast that aims to educate and inspire music creators in their quest to achieving their goals by gaining a greater understanding of the business of music. A new episode is released each Wednesday and aims to offer clarity and insight into a range of subjects across the music industry, all through the lens of a music creator for the benefit of other music creators. The series includes soundbites and interviews with guests from all over the world together with commentary and clarity on a range of topics. The podcast is hosted by award winning music industry professional Jonny Amos.
Jonny Amos is a music producer with credits on a range of major and independent labels, a songwriter with chart success in Europe and Asia, a senior lecturer in both music creation and music business at BIMM University UK, director of The SongLab Ltd and the author of The Music Business for Music Creators.
www.jonnyamos.com
The Music Business Buddy
Episode 33: An Interview with Gary Charles on Music Evolution and AI
Join me as I sit down with the innovative Gary Charles, whose journey from post-punk indie band member in late apartheid South Africa to electronic music pioneer offers a unique window into the evolution of music creativity. Gary's story is a compelling narrative of transformation—crossing from the analog world of traditional instruments to the digital realm of electronic sound-making tools. Those interested in the groundbreaking work of Aphex Twin or the LA beat scene will find Gary's approach to music production both fascinating and inspiring.
Gary shares his experiences navigating the vibrant techno scene in Berlin and the challenges of transitioning from band life to electronic music creation, using early software tools like Logic and Reason. This episode touches on his deep dive into sound art and his innovative use of AI and machine learning in music production—a passion that eventually led him to pursue a PhD. Beyond his personal projects, Gary is committed to teaching and collaboration, helping others discover their unique soundscapes and remixes.
Our conversation also grapples with the ethical implications of AI's role in music, examining the risk of cultural appropriation and the biases inherent in AI models. We question whether AI's ability to recreate music might stifle innovation and overlook cultural context. Gary provides insights into the monopolistic tendencies within the music industry and stresses the importance of community and cultural connections in music-making. This episode is a must-listen for those curious about the intersection of technology and cultural heritage in the modern music landscape.
Websites
www.jonnyamos.com
https://themusicbusinessbuddy.buzzsprout.com
Instagram
https://www.instagram.com/themusicbusinessbuddypodcast/
https://www.instagram.com/jonny_amos/
Email
jonnyamos@me.com
The Music Business Buddy. The Music Business Buddy Hello everybody and a very warm welcome to you. You're listening to the Music Business Buddy with me, johnny Amos, podcasting out of Birmingham in England. I am the author of the book the Music Business for Music Creators, available in hardback, paperback and ebook format. I am a music creator with credits on a variety of major and indie labels, as a writer, producer, and I'm a senior lecturer in music business and music creation. Wherever you are, whatever you do, consider yourself welcome to this podcast and to a part of the community around it. I'm here to try and educate and inspire music creators from all over the world in their quest to achieving their goals by gaining a greater understanding of the business of music.
Speaker 1:Ok, so in today's episode I am interviewing Gary Charles, artist and music producer, exhibiting work including moving image, sound art and performance. Alongside extensive recording, mixing and mastering experience. Gary's also a senior lecturer in music production and has composed sound and music for artists, film, theatre, contemporary dance productions and many other things. Another facet of Gary's career which is particularly interesting, and something I'll be asking him about today, is that he is an Associate Practitioner of the Sonic Art Research Unit at Oxford Brookes University and is currently undertaking his PhD research looking into the impact of artificial intelligence on creative practice and cultural production. So let's get into it, gary Charles. Welcome to the Music business buddy. It's good to have you here. Thank you for being here. How are you, mate?
Speaker 2:I'm great Thanks for inviting me. I'm looking forward to it.
Speaker 1:Oh good, I'm glad you're here, right? There's so many things we can talk about, gary, but let's start with you as a music creator, right? So some people will know you as the static hand, other people will know you as the guy from Low Red Moon, others will know you from your work with Spectral Karaoke. But being a creator and a producer defines you and your music career. It's kind of it sums up you as to all the other things that you've done since and before that. It defines you, right? So how has your creativity evolved during your career so far?
Speaker 2:I started out really in post-punk indie type bands as a teenager. I grew up in late apartheid South Africa where I guess they were more interested in teaching you how to shoot a gun than play a guitar or play the piano, and so I never had a. I never had a musical education. I don't come from a musical family or anything like that. It really just just came from some, from being really into um, into the music, and started getting involved in in different. I couldn't play an instrument. I still today can't play any specific instrument particularly well.
Speaker 2:I mean, on all of my records there's keys, synthesizers, guitars, bass that I play, but I'm not sort of classically trained at any of those things. That's very modest, gary. So I mean I sort of every time I've tried to also learn in the classical sense, there's a little trigger in me that says, no, you just want to go and make some noise. So I tend to approach most musical instruments very, I guess, and musically and and and start to use them as sound making uh tools rather than than um, thinking about, you know, thinking about notes or chord structures, which is yeah, which is just something that, yeah is kind of the way the, the way that I've worked, so I've I've tended.
Speaker 1:That's quite punk in itself, though, actually isn't it?
Speaker 2:that's well or it could be just laziness.
Speaker 1:No, it's definitely not laziness. I know you, you're not lazy so, um, so.
Speaker 2:So I was in sort of, yeah, uh, post-punk indie we were into, we were into sort of, um, I guess, artists like the, the stone roses, happy mondays, also I guess, a bit of cure suzy and the banshees, uh, the fall bands, bands like that which, um in sort of uh, early 90s south africa there was. There was not many people, you couldn't really hear any of that stuff on on the radio um, but one of the one of the guitarists that I was, um playing in a band with his brother, had a basic recording setup um at home and he also had a uh, a sort of old one of those old, really kind of basic elesis drum machines and so I started, great, those things started, yeah, programming, programming those and really really enjoying it and um, he also had a um I can't remember what synth um he had. It was a sort of one of the um nord nord type type synths. So I started kind of playing, playing with that and and and really kind of getting into um, getting into the recording process as well.
Speaker 2:So I got we were talking a little bit earlier about the the porter studio cassette, oh yeah, so I got one of those and we should have recorded that yeah, I love, I love those, I love those um sort of old four track cassette, cassette recorders um you can still find them as well when you look for them on ebay and you, you pay a fortune. Yeah, especially the ones with the, the sort of of pitch control or the, the the time oh right, yeah, they're harder to find, yeah, you can almost use them as an instrument in themselves, yeah, great.
Speaker 2:So so, and at the same time I started getting into into more electronic music, so things you know, things like fx, twin or tecker, um, uh, yeah, um, some of the. I guess when, when the flying lotus and the sort of la la beat scene um started becoming prominent, I was kind of really into that, so started started becoming more electronic focused um, also recording some demos for for other other local local bands, and so really kind of started getting into the, getting into the um production side, I guess the recording and engineering and production side, but entirely self-taught and I didn't have any money, so so it was kind of like um teaching and and scrabbling around for whatever we can, you know, get our, get our hands on um, and then then um, I decided I decided after um, after studying actually I studied law because I kind of was really first.
Speaker 2:Yeah, wow, which was? I never wanted to be a lawyer, but it was kind of one of the most rewarding things because it was at a time when the constitution was changing in South Africa so to just become a democratic country, changing in South Africa so to just become a democratic country, and a lot of my professors were the advisors to the ANC in building this new constitution for a new country. So really, really a fascinating and wonderful time to be studying law. So I kind of sat in some of the early constitutional court sessions where they, you know, overturned capital punishment and all all big sort of, big sort of decisions like that, so that was really wow, that's amazing.
Speaker 2:And it was kind of a privilege to and a lot of my old professors are still very active today.
Speaker 2:So, for example, john Dugard, who was my international law professor he was recently the head of the team that brought the case to the South African case, to the IC, icj, to the International Court of Justice um, uh, so, so it's kind of a lot of these people, even though I was just there as a, as a student of, it's kind of stayed, I think, really, um, yeah, really really important for me. But at the the same time I was, I was kind of playing in bands and wanting to get deeper and deeper into into music and I decided to come, come back to the back to the UK, but in in the meantime, what I did was take, take a couple of months off and spend it every penny that I'd saved up, and I uh took two months to just learn how to. So I I'd already had a uh sort of I guess uh cracked version of of early logic when it was still pc and mac. Um, well, e-magic gosh was.
Speaker 2:So I'm showing my age all right wow okay, jeez, um, and then I, and I got reason, the, the first edition of of reason, which had the. When you flipped the panel it had all of the, the, the cabling, yes I remember that like a virtual virtual analog where you had to plug everything into the mixer.
Speaker 1:Yes, did you ever like hit the tab button really quickly to try and make the cable shake a bit more? Yeah, exactly.
Speaker 2:Great, and it was just such a for somebody who didn't have access to, you know, studios and any of the equipment. It was such a brilliant way to learn and so I kind of learnt first in the virtual world, in the early, early versions of Reason, so I mean, and then eventually I started using Ableton Also. Ableton won the very first oh right, wow, I'm that old.
Speaker 2:Crikey, you've paid your dues man, wow, I'm that old crikey. You paid your dues, man, so anyway. So, so, so, um, um really learning all along, and then started started doing um making some, some stuff on my own as a, as a solo artist. So the static hand is probably the the sort of um project that that's been the longest standing and has been around for a while. But I've released music, all sorts of weird and wonderful things, some of which I wouldn't probably stand by today or want to listen to, but under all sorts of different names, but mainly focused, more focused on sort of ambient, experimental electronic. And then some of the projects were sort of more, I guess, dance floor focused, return to the UK for a while and then move to move to Berlin.
Speaker 2:So well, in lived in London and my um partner at the time, who's now, who's now my wife? Um got accepted to the Royal Academy and she's she's South African uh got accepted to the Royal Academy of Art to to study, to study um fine art. So we kind of made a pact study to study um fine art. So we kind of made a pact um that I'd still go and work and do mundane office jobs to to kind of get us through the, the three years, and then after that we will both just be artists, musicians, um, and and go find somewhere that's affordable to to live, which we did. And we we went to Berlin, um, and so then I became much more kind of involved in the, the, the sort of techno um, electronica kind of kind of scene. Um was in a, was in a, I guess, electro punk uh group for a short while with a um, um, which was a sort of a short-lived project but quite quite fun, and then also started mixing and mastering um tracks for uh, essentially for for djs who who uh, because the the club scene was was kind of kind of huge there and there were lots of djs who were looking to to learn how to produce and to get their records to sound good, not so much on cd or at home listening but on the, the, the club sound system where they were, where they were DJing and and and in most cases where we're resident DJs. So so I kind of got involved, involved in that.
Speaker 2:I had a little studio in the old communist East German national broadcasting center Wow, the Funkhaus Really Good job, I've heard about that Wow which was just great.
Speaker 2:So my wife had a large painting studio downstairs and I had a small little recording and mixing and mastering space. I also started then teaching which was really new for me and generally teaching DJs who wanted to learn how to produce. So I'd kind of be with them all the way through, kind of creating their first records and trying to help them make sure that it sounded the way that they wanted to sound on in their sets and and that was like that was just really good fun and um, and at the same time I was kind of I was releasing um, uh, you know, under the different projects that I was, I was I was working on releasing um, releasing tracks, did a bit of remix work for florian meindel, who was who was um, kind of a fairly well-known tech house. I mean, all of the all of the scenes that I've been involved in up until this time, or always, really were kind of underground, I guess.
Speaker 1:Yeah, yeah, kind of underground niche, Well, which is where things are at again now. Right, you know, with Discord servers and communities, and you know micro genres, I mean that's a lot of different scenes as well that you've been a part of Gary. That's a lot of different cultures and styles, all kind of mixed there. I mean you've also done stuff for film and theatre and contemporary dance productions, so you've done all sorts of different things. Is AI a part of your workflow now? You know, is it integrated into the way that you do things now?
Speaker 2:So the AI thing really came about through well, through fascination with technology firstly, and then, when we came back to the UK, I did a master's at Oxford Brookes, at their, in sound art, because I'd started doing sort of sound art, installations, artist films and sort of more, focused on the sort of I guess, the art school world rather than than the, the, the typical music, um and music industry kind of world and um, and I decided to do a, I decided to do a, a PhD, and at the time um, I was starting to work with, with early machine learning models. At that point in time, if you said to anyone I'm going to do something in in AI, people didn't really have a, have a clue what you're talking about. And now flash forward sort of six years, seven years later and you know it's, it's kind of everywhere and I'm still busy doing my, doing my phd on it, which is which presents a number of challenges. But um, it started out where I was incorporating um, ai, early, early sort of um, midi generation models, um, some the, the, the kind of early audio, short clips of audio, um generation models and um. I started out using, using them and incorporating them into my practice, um and into the, into the work I was doing.
Speaker 2:But the more and more I got into or delved into this world, the more my thesis became a critique of what is happening and and the emergence of AI, not just in music kind of in general. So, and I've done a few sort of multidisciplinary projects, I guess you'd call them, kind of looking at that. So the last one is the last one that we, so we were invited to do a residency in in South Africa. So myself and my partner, who's my wife, who's a visual artist, mainly South African poet and playwright, and a um, who's Napo Mashiani, and then Tuli Sile, who's a um, a dancer, a performance artist, and for each of the, for each of the, the disciplines, I guess you you'd say we, we worked with the latest uh sort of AI model to interact with them and kind of try and break them.
Speaker 2:Essentially that's what my practice with ai has been trying to break okay, these the, the, the models, um at the time and see kind of how they work and and understand them and in working in South Africa, is really interesting from that point of view because of course, the training sets for, even for the sort of very large language models like ChatGPT, are really based on scraped information from the internet and are therefore heavily Western biased.
Speaker 2:So scraped English content from the internet, heavily Western biased, heavily US and European biased. So when you're working with things like so, for example, working with things like so, for example, one of the first um, uh ai models that really kind of I guess uh passed or or could be seen as passing the turing test of convincing somebody that um, it was actually um, the, the, the work of a professional human, was Die Bach, which was a model trained on the Bach chorales and if played to a non-trained listener, so a lay person, they heard one of the Bach chorales generated by this AI model. Essentially it took all and of course both the Bach chorales already have a pattern and an algorithm that was embedded in the, in the music, and so people wouldn't be able to tell which was an original Bach chorale and which was a um which was AI generated right.
Speaker 1:Okay, that now, gary, that right. There is the problem, isn't it? For a lot of people, especially, you know, rights holders in the music business, right? Is that like? This is great, but what's it learning from? And if we know what it's learning from, it's like why isn't that remunerated? How could it be remunerated? A lot of questions around that that we don't have the answers to yet as time ticks on. But in many ways it's got. Language models are kind of the way that they scrape there, right, it's like I'd never thought about the bias aspect of it before, right, but of course. Of course it's bias, isn't it? If it's just, if it's just web scraping and running paradigmatic analysis in music it gets.
Speaker 2:Even in music, I think it gets even more um, complex when, when you think about the, the, the from from a music industry or western popular music point of view, the first question we we have and you know kind of quite quite rightly is what's in the training set, what, what do the copyright holders have and do the copyright holders get remunerated and that that sort of. You know, we've been here before with things like Muzak and with Napster, where we're kind of trying to uncover these things, and that's an area that I'm obviously super interested in. But I kind of zoom out and see a slightly different problem. Like, I kind of zoom out and see a slightly different problem because if we manage to get that resolved, which you know, there's big court cases between the major labels and UDO and Suno, the prominent AI generators at the moment, and so it can appear that that's kind of maybe going to be resolved. But when you think about, say, amapiano or Qom music, something that's very culturally embedded in a location where in South Africa, if you're a young um or I'm a piano producer, you'll be creating tracks, usually in your bedroom with a cracked version of FL Studio or Ableton or whatever you can get your hands on and that music is really even even in south africa. A lot of the, the, the kind of genres of, of music emerge from certain regions, um, and they're very culturally embedded. What are we saying if we say, well, actually we've solved the, the problem of ai and copyright, because we've had the court cases, the major labels have have done a deal that they'll say is is protecting artists, um, and then we say, okay, that's solved, so now we can create a. I'm a piano generator, um, so.
Speaker 2:So then you've got a, the silicon valley monopolies and, uh, the music industry monopolies, kind of colluding to essentially appropriate the entire entire cultural underpinning of of a music that's very embedded in in communities or in in in locations. And so then suddenly you'll have a, um, yeah, a silicon valley tech startup, you know, kind of um, which is reliant on the infrastructure that is essentially monopolistic for chipsets. So, at the risk of getting sort of mundane, but in order to train a model you need not the CPUs that are in your computer but a graphic processing unit, and really the hardware itself is already a monopoly. So, which is why nvidia has become, um, one of the world's largest companies over the last, you know, last uh, five, ten years um, as opposed to just being a producer of high-end graphic processing units for video gamers, um, they're now the, the, the processing unit for for all of the servers that um that are used to train ai models. So so, basically, the entire infrastructure is, is, is a monopoly um, and if we accept that, okay, we, we've, we've, we've, we're, we're, we're fine that that uh, copyright holders will be remunerated in places like South Africa, or, if you think of um, uh by funk uh in in Brazil, or you know uh, anywhere where there's there's there's kind of musical innovation happening and that can easily be just just transported across and there's not really um as well established copyrights, copyright holding function in in in those countries.
Speaker 2:So so, for example, in south africa, if you've made your your track, usually what you'll do is um, put it on a usb stick. Well, these days, um uh, send it, send it off to to uh taxi owners, and, and the taxis that that transport people to work every day are one of the, the places where people hear music and where so, so, for example, here in um, if we think're kind of often seen as, as as niche, um, niche, exotic types of of music. So, for example, um, uh, I saw shoma josie, who's a uh in south africa. She's a pop star, she's like the south african beyonce, but uh, when she's playing in europe. So I saw her at the CTM festival, which is like an underground electronic type festival in in in Germany. Um, so these, these, these types of music are seen as kind of niche and you know, and there's not not much copyright going on.
Speaker 2:But then what we're saying is is that that that entire cultural lineage is open to just being imputed into a, into a model.
Speaker 2:And I think that's that's where where I kind of have have um, the, yeah, the concern or or the, the area that I, that I kind of like to to look at, and then in in in the language models in chat, gpt, I mean, even today sort of, if you, if you kind of kind of try invoke cultural references from, um, you know, american or or brit, british popular culture, it's usually got a lot of information to draw on.
Speaker 2:But if you start kind of questioning about the life of a contemporary, a Sissutu person living in Johannesburg, it really doesn't have any of that context and it kind of ends up drawing on stereotypes from sort of Western culture. So it's kind of wild, like Dali and Midjourney, the image generating models. The image generating models even today. If you, if you kind of ask it to give you an image of a wealthy, a wealthy Zulu banker, for example, it'll often, or and their home, it'll often show a man in a suit outside a sort of mud hut or you know kind of a uh. So it's still still got those those stereotypes and and biases deeply embedded in the, in the, the training sets so it's looking through a westernized lens?
Speaker 1:yeah, absolutely. So what? Um? This is a really, really, really interesting subject, gary. Thank you for bringing it up. I appreciate it. I mean, um, one of the things that's at the top of the agenda for um ai is is is what I mentioned earlier right, rights holders, you know, etc. And there may well be a path to stabilising that at some point. But what you've raised is another ethical problem. So let me phrase it this way, let me frame it in this sense right, if let's take Amapiano, for example. Right, if we've got, say, say, five or six artists, um, that it's amapiano, it's, it's, uh, it's from originates from west africa, right, ghana, nigeria or no, no, from from south africa does it.
Speaker 2:Yeah, and even even within south africa there's there's sort of contention of where, where the lineage. Because you obviously had Kwaito, sort of post-apartheid South Africa. Kwaito was very popular sort of dance, pop music, and then you had a lot of genres springing out in Pretoria and now Pretoria to give you a sense of even how local some of these discussions can be Pretoria is, I think, 30 odd miles from Johannesburg, but Pretoria there's a genre called Bacardi House that was very popular and that emerged there, and the Pretoria folks will tell you that Amapiano is from Pretoria and Joburg folks will will tell you that that Amapiano is actually from Joburg and comes from a slightly different lineage. So even within South Africa there's a regional dispute and Qom, for example, is from Natal, and so even within South Africa you have these regional disputes.
Speaker 2:But then when we're looking at these, these kind of genres, and they all have a a regional and a cultural lineage, that kind of um, that that goes, goes back through those and and that's kind of invisible, that context is sort of invisible to us, so that in fact we get, you know, kind of the idea that kind of Africa is just this one place creating music, whereas it's kind of hugely multi yeah, multifaceted.
Speaker 2:And which also leads me to another kind of challenge with with AI, recreation or reproduction. If we also get used to the fact that that and accept the fact that AI, this technology, is so well developed that it can create passable, let's say, I'm a piano and then we can produce that music at scale always, that's looking back, we're just reproducing. We're just reproducing the past, because every time a training model or a model is trained um, it's only got historic information and then we kind of reduce the ability for new innovation so that that, like organic shift from kwito to um, to korm to shangro, to all of these regional takes that emerge. If at the point let's say 1999, we say we build a model now of South African pop music and we can generate K a for days, then would would all of these, these new genres or new modes of music making of, of have emerged or have strongly emerged.
Speaker 1:So it's a well, that's a good question, yeah, and especially in a world where you could take a producer who could sit in I don't know on his laptop in Peru, right, and pull in, let's say, some Amapiano logarithms, right, and then combine that with, let's say, a Norwegian folk rhythm and pull that together, and then, all of a sudden, the scene starts where he or she or they are based and all of a sudden it becomes basically a fusion of everything that's come before. That's a good example of AI, isn't it? It's like it is pulling everything from the past, but the pieces, when collected, might build something new. I don't know, am I looking at it too positively? Tell me.
Speaker 2:Well, I mean the thing I always always kind of, because we already do that right, like so, so to to some extent, but to to, for that to be interesting, I would argue it needs a cultural movement around it. So these, these things are, are not simply embedded in the notes and the, the, the notes and the rhythms. They can't they, they can't be, uh, they can't be kind of encoded easily. So so, for example, um, when, when these, the, these kind of movements happen, they're usually based around a region, sometimes individual clubs. Think of techno in Germany with Burghain.
Speaker 2:Think of house, music comes from the name of a place, a club in Chicago. So scenes and cultures aren't just the music and the, the notes that are, that are played. I mean, I always think of, I always think of, um, uh, vudu ray, a guy called gerald, so it just embedded in that music, so is it makes you think of sort of post Thatcher Britain, the Hacienda, uh, the technology. So the cheap 303 808 then were kind of uh, classified as a, as a complete, you know, um, failure, and these are suddenly making the biggest sounds. And obviously you think of ecstasy, you think of acid house, so all of these cultural connections that are brought to bear on one piece of music if that sits in a training set. All that we have is the 48,000 digital samples per second, so that entire cultural context gets stripped away from the music.
Speaker 2:And so if we see music as simply the notes and maybe this is this is a grand way of me me justifying not learning, not learning how to play the piano properly, but um, but if we, if we start to reduce music to, we're just creating a commodity, I think it's deeply problematic. And the logics behind ai generation, I see, is kind of that logic on hyperspeed, like the aim of music is just to create, create this, this wav file and and really then that removes all of the, the sort of musicking people. Music is a social action. Uh, music is a cultural action. Oh, just dropped my phone, sorry about that.
Speaker 1:I was getting, I was getting a little bit amplified, I was about to get up and start toy-toying, but there's some really, really good points here. Let's take a couple of examples. First off, if you take a tradition that, let's say, is difficult to actually say where it's from you mentioned about Amapiano and the different arguments around that, for example. So if something disappears into the mists of time, at what point do we start to regulate its foundation? And then what does that look like when it is learned from? So if we split into two categories and we go right, one's tombra and one's mathematics, right, one's patterns and one's tombra, then we get somewhere, don't we? Because we can say say, well, we can take this particular rhythm from that particular area of south africa. But if we do it in midi and we apply it to, I don't know, a nashville gretch kit, then we've got something which is a hybrid.
Speaker 1:It could be argued that that's not new, but it does come from somewhere. That should be respected. But, on the other hand, if we're just taking the timbres and sampling them and using them differently, again that becomes something new, but based solely on the origins of that original source. Yeah, so there's, there's, that's okay, let's think about this one then. Go right, you mentioned, you mentioned about playing piano there, right?
Speaker 1:I'm a piano too, piano Now. Okay. Now some people get really they don't like it when we talk about these things, right? So I don't mean to annoy anybody, but if we take out the wholesome aspects of playing an instrument and we just think about it for the purpose that it is to record and perform, right, let's just put recorded music under the microscope for a minute.
Speaker 1:If you didn't know how to play the piano, that doesn't mean obviously you can't have piano on your record. Somebody else could play it. We know this, that's been around for hundreds of years. However, if you were to be able to play piano and you wanted to record that onto a song that you were creating, it's going to be a little bit biased on some level, right, because it's you and you like it, because it's you playing it, and that's okay because that's where art comes from.
Speaker 1:However, there's a limit to what you might feel on that particular day, how you're conditioned, how tired you are, etc. But if you're generating patterns on that piano on whichever piano let's say it's a contact instrument or whatever and you just keep generating patterns, the thing that dictates whether you keep it or not is the decision that you make, where you say that's the part I'm going to use it Now. It could be that actually that yields and for many already does yield a far greater result than having played that instrument in the first place, which is what I mean about offending those that are wholesome instrument players, for which I am one, right with guitar. I've played guitar for like 30 years something, but very often I choose samples because it's quicker to work with. Shoot me down, I'm sorry. So you know.
Speaker 2:It's like you're in very good company with with yeah, I think I am.
Speaker 1:I just don't want to offend any listeners, but you know, it's a thing, it's here, but it's like when we really really break it down, it's like it's laden with innocence in many ways. Yeah, because you know, if we were to take a particular sound on Splice and say, right, let's put that in my sampler and build something with it, it's like we should have taken, we should be upon ourselves to understand where that came from. But it's difficult to know. I mean, that's we can't. We can't regulate, that can we.
Speaker 1:But but how do we like, if we, if we take a genre, for example, like one of the ones from south africa that you mentioned, that you go right, yeah, if we are doing something unethical here by taking a genre, if that genre had sort of five or six or flagship artists, the previous work, the manner that we worked out in terms of copyright collection, would work, wouldn't it? Because we'd be able to say, right, those five or six producers or creators of music are synonymously linked with that style and we can use that to train the model that we're going to use to recreate it, in which case we can remunerate those five or six people, but not the genre or the place it came from, right, so then it does work. So then it's like well, the argument we could become, well, you know, we can train a model to understand queen, yeah, the band, but we're not taking into consideration the origins of where they took their influences from. It's like where does it stop?
Speaker 2:you know what I mean the way that I sort of see it, and and this is not just a, an ai problem, but it's a, I suppose, a capital problem I mean the, the music industry, the popular music industry, or the industry with a capital I is already a monopoly and I think that's problematic. So in that scenario, the AI industry, all the way from the infrastructure that is used, is a monopoly or monopolistic, and so the challenge with that is you've got everything controlled by very few hands, so that ability for younger or somebody doing something interesting and new to come through becomes reduced, which is already to some extent the case. Seeing it, I guess, guess from a positive perspective, and I mean this this is a challenge that we, you know that there's a lot of work being done with with spotify. Spotify, as you know, is um part owned by the, by the large record labels, um, in a reaction to and and in a reaction to to the threat of napster and peer-to-peer um pirating, at the time, that was was done to make sure that the, the, the major labels would tell you it was in order to protect the artists. But the major labels get a minimum, minimum value per month from spotify regardless, and our shareholders, so that whole system is really designed to, to control the entire industry.
Speaker 2:So so from that perspective, I think the music industry is a as a whole, is, is is problematic from from my lens, just from the point of view that it's, that it's so monopolistic, but then interacting with another monopolistic system, the tech, the, the tech universe, uh, I don't see how much good can come of it. I don't really know what, what to do about it or what I mean. I guess, from my perspective, the, the, the approach is to to try and instill in, in, in in people, the, the joy of community, musicking, together, community building and and culture building. Um, but it is, it is kind of a challenge if you are wanting to make music your, your career yeah, well, that's what I wanted to ask you.
Speaker 1:That's what I wanted to finish off with. Was that question right? Was that like what's the advice that you could give to emergent music composers and producers in regard to their relationship with with ai the?
Speaker 2:the weird thing is and maybe to leave it on a, I don't know if it's a positive, but an interesting personal kind of take. I mean, as I mentioned, most of the music making that I've been involved in has been kind of in niche scenes and I've always sort of ignored pop music, as you know, uninteresting or not for me. But more recently I've become much more fascinated with pop music production and I think it relates to the AI and there's a number of reasons for it. You know, working with some of the students who are so super talented and are making their way in this world and are really gifted pop songwriters, producers, that's kind of got me a bit more interested. But also, I think the weirder things are coming through pop music production and I do think it's related to AI in the sense that if AI and not just AI, even sort of the immense amount of tools we have that allow us to do everything in key, to quantize every bit of information, to sing key, Essentially it's easier to make music that's right than music that's wrong at the moment. And if you know how to use the tools and in that world where it's easy to reproduce the generic, then maybe the slightly weirder, become popular. So I mean, for example, the Charlie XCX record.
Speaker 2:Ag Cook, the producer you know, comes from a real sort of left field underground scene. He started pc music. He did a degree in music computing or in computing at at goldsmiths, so really a really interesting producer and has produced probably the, the record that's going to win the gram for the best pop record for last year. That's interesting.
Speaker 2:And then you've got somebody like Lewis Roberts who produced the FKA Twigs record. He releases music as coreless Really interesting dance music without beats was one of the terms that people talk about him. So people with interesting takes and doing new things are more involved in the pop world. So that's kind of fascinating to me and I and I hope that we see more of that yeah, I think, I think, we, we probably will and it all creates more learning, right, more learning that can be trained.
Speaker 1:Uh, no, uh, that's uh, that's fascinating. No, no, it's some. You know I've said it before on this podcast, gary, sometimes the questions are more interesting than the answers. Of course, you know it's good for us to talk about this. Your insight on this, gary well, on many, many things, is fascinating, and I really, really am grateful to you for coming on here and talking about you know about your background, about all these different things that you've been through musically, and how it informs you to talk about ai and and its role, and especially, you know, some of the ethical values that don't get talked about as much when it comes to ai. So, uh, thank you for being here, mate. It's much, much, much appreciated. It's been an absolute pleasure. Thanks for having me.
Speaker 1:Oh, what a super dude don't. I get some lovely people on this podcast. I don't mean that as a big up to myself. I just mean there's so many great people isn't there in this world and in this industry, and I can't even begin to explain how much I learn from just not talking to them but just listening to them, and Gary's yet another example of that. I mean he just has such a multi-dimensional perspective on things because he's lived through various music scenes in different parts of the world and it just, you know, it doesn't give him a purer soul, but it doesn't half make him more useful as a music creator. He's just so well informed.
Speaker 1:And I will say it's kind of nice to change uh, you know, to switch up the subjects and the angles on ai a little bit, because it's very easy to just buy into the very kind of social media driven narratives on, oh, isn't this terrible? Or oh, it's replacing humans, or, you know, any of the nonsensical things which will prove to be probably rubbish in various years to come. And it's also the agendas that we hear on the business side of music, which I can understand in regards to, you know, what are we going to do to remunerate rights holders? These are all good questions. There is a lot more to AI than just those things, and those are the things that you know Gary kind of talked about here today and, yeah, it was really really, really interesting. I hope you gained a lot out of it. Ok, I will see you again next time. Until then, may the Force be with you.
People on this episode
Podcasts we love
Check out these other fine podcasts recommended by us, not an algorithm.
The New Music Business with Ari Herstand
Ari's Take
Music Industry Podcast
Music Industry Podcast
The Music Education Podcast
Chris Woods
The Music Industry Podcast
Burstimo
Music Business Insider Podcast
Ritch Esra & Eric Knight
Inside The Music Business
The Red Cat Agency GmbH
Music Business Worldwide
Music Business Worldwide (MBW)