Your AI Injection

AI In Music Generation: Interview With Jay Bartot

July 20, 2021 Deep Season 1 Episode 6
Your AI Injection
AI In Music Generation: Interview With Jay Bartot
Show Notes Transcript

This episode of Your AI Injection features Jay Bartot, a tech entrepreneur and musician. We chatted with Jay about his upbringing in the Chicago music scene and the expanding world of AI-driven music composition.

Music is central to modern entertainment and culture. It serves as stand-alone art, as a backdrop to daily life, and as an important feature in industries like film and advertising. But
just as work by writers regularly appears alongside robot authors in sports, financial, and other news, human musical artists are increasingly sharing the stage with artificial intelligence. What if the music you hear on the radio was not composed by a human at all, but by a machine?

Check out the article on our website about this topic!

Automated Transcript

DEEP: Welcome back to your AI injection. This week's episode features, Jay Bartot, Jay's a technology, entrepreneur with a number of startups and and a few nice exits as well as he's also an innovator Jay's, got 20 plus years of experience, developing data and machine learning apps for businesses and consumers. Jay, thanks a ton for being here.


JAY: Hey deep, thanks for having me.


DEEP: Cool. So we're actually gonna be saving, I'd love to sort of dig in with you about your Tech background and you know, just for our listeners benefit, the two of us have worked together on multiple projects over the years and J's. You know, everyone's got that short list of people, they just love to work with. And J, certainly, one of mine if not at the very, very top of that list. So we'll save that Tech entrepreneur conversation for later. But we're going to talk about music today. So, Jay and I, we've nerded out on AI systems for years, but we also have had the privilege of being able to play some music together. He's got one of the most amazing music studios that I've had the privilege of being a part of. And so with that Jay, I want you to go way back and tell me about your, like, how you got into music? I remember you having a pretty awesome story. How you got into the punk scene? Take us back into into those days?


JAY: Yeah. No happy to. And, and I appreciate the sentiment on working together. You're at the top of my list as well, and so hope to see us keep doing that. So so yeah, so you know, my music background in my technology and Entrepreneurship career are actually pretty closely linked and that will be obvious as I start to tell my story. But you know when I was a kid around, nine years old, my family moved from rural Connecticut to Inner City Chicago. My mother was pursuing a PhD in sociology at the University of Chicago and and so I had three brothers and sisters and a dad, and she picked the family up and moved us in the late 70s, and plunked us down in Hyde Park in Chicago. Where the University of Chicago is and as you know, deep you and I have connected on this before. Yeah. The south side of Chicago is a little bit. It's a little


DEEP: rough. Yeah, I know for our listeners benefit. I went to school not too far from from U of C at the Illinois Institute of Technology. And, you know, to if anyone's read Freakonomics, they are very original version. That was pretty much our life. We were surrounded by the heart of the North American crack trade back then. So it was, it was rough? Yeah, yeah, yeah.


JAY: So you know II had certainly build up survival skills and Street skills and and was, you know, plunged into that world, you know, pretty dramatically. And so, you know, I was a pretty nervous anxious kid and and so not long after we arrived in Chicago, I had to pick an instrument at school because, you know, being in fourth grade I think you know that's that's when those Decisions came. And and and so I picked the drums that summer that we had moved from Connecticut to Chicago was, you know, I think in retrospect, you know, pretty anxiety-provoking and I taken taken to tapping and beating and as my siblings used to say, banging on any surface. I could find looking for an interesting sound or tone. And so, so I took up the drums in the school band. And you know, got into it pretty quickly. At the same time I had older siblings who were, you know, into the 70s rock music of the day and so I was probably more of a sophisticated, you know, listener, you know to, you know, rock and pop music up that that era then say you know some of my classmates were given I had older siblings and so you know it didn't take me long. Long, you know, to be playing my drums along to the radio and and all the bands of the time, you know, of course, is the 70s. So, you know, we were all listening to Led Zeppelin and and Pink Floyd and and all the 70s rock bands. And then in the early 80s, you know, bands like the police in Genesis and so forth. You know, we're certainly things things that I listen to And by the time I was in eighth grade, I had formed my first band and gathered up a few few other kids in school who are playing guitar and bass and singing. And, and we had, we had our first band and because we weren't very good. I thought, you know, we should play Ramones


DEEP: covers. You got the remote one string bass line, and yeah, totally. That's that makes


JAY: sense. Yeah. So we were playing Blitzkrieg Bop and you know, a variety of other Ramones tunes and and, and starting to play, you know, some some gigs had, you know, in school, at some at lunch time and a few parties here and there. And so, you know, but by the time I was going into high school, I was convinced that I wanted to be a professional musician and be a professional drummer and No, that was what I was going to do. And so I spent most of High School playing in bands and you know, I pretty quickly graduated from playing with kids, my own age to playing with grownups and adults who are older than me that that they were in their 20s. And and the main reason for doing that wasn't because I was so good or talented or anything. Certainly, I had lots of schoolmates who were who were talented but I just Found that older older people were more serious and focused and you know, kids my age and high school. It sure they were playing in a band but you know that that wasn't necessarily their top priority. And so I was very passionate and focused and so spending spending a lot of time doing that you know, one of the things I noticed early on in early high school was that You know, if, if you were into 70s rock and and 80s rock, you know, there were, that's where the masses were, that's where the kids who, you know, again could play. But there weren't really any gigs, you know, to be found, you know, plan Led Zeppelin covers and and so forth. Yeah, you can play a school party or something here and there, but I got introduced to hardcore punk in the early 80s in Chicago. And what struck me about it was that it was a scene, you know, it wasn't a bunch of kids who, you know, we're just trying to do their best to imitate Jimmy Page, it was very raw music. It was playable music much, like the Ramones that I had played in. First band, but, you know, they were clubs professional clubs where these gigs were happening. And and there was even like a local Chicago hardcore, punk fanzine. And so, there was a, there was a scene happening. And I remember telling people that once I joined one of these bands that, you know, I was playing at, you know, such and such Club. There was a club called Tut's on the north side. I think I'm trying to remember exactly where it was but you know in the near Northside that was a well-known music club and you know they were trying to make money like everyone else and so once a week they would have hardcore you know Punk night and all the punks in the Mohawk stuff would come in. Yeah. And and so I was playing in real clubs by the time I was 15


DEEP: so that's kind of Be like one of the things about that's that's intriguing and I'm sure you're going to help us make this transition. But but Punk is sort of special for its spare Earnest and rottenness and it kind of feels like nothing could be more not Punk than like, computer music and and, and sort of, you know, all this digital, you know, who are whatever, you know, like punks pretty raw. And so, when I was back in Chicago, I was totally obsessed with Ministry. And you know in these guys are, you know, they're there, I don't know. They're like kind of like maybe Punky evolved into like, you know, super fast industrial for those who don't know. And one of the things that struck me about Ministry was they had these like crazy drum machines that would just play a lot faster and tighter than humans. Really did back then. And one of the things that I noticed is similarly like a few years later like drum and bass had a similar phenomenon where people were like work. In this electronic World there, you know, putting together these drum lines that normal humans didn't really play at the time. And then you fast forward a few years and even like Ministry themselves had a live drummer, they started playing this stuff. So they started playing like, quote robots or like you know, like the machines were playing and same thing with drum and bass. Like they became this whole thing where, you know, it's just spare, its the snare, it's the high hat and, you know, and the kick and you're you're copying machine. So the question for you is like, are you seeing This kind of thing even today, you know, this sort of where the machines are teaching the humans. And the humans are like, like it's this this, this feedback loop where it's not just replacement. It's more like almost inspiration from the other side. Yeah, great question.


JAY: I'm not sure how much of that I've seen, you know, certainly in all, you know, pop music. There's always an element or an edge of, you know, people who are doing things that are Quasi, you know, virtuoso or acrobatic. And so, you know, I think about metal, for example, and, you know, watching These these these bands come up with, you know, you know, incredible chops and you know, people practicing, you know, hours and hours a day, probably, you know, similar to, you know, what, classical musicians, you know, we're doing during their training the long hours and putting the passion and effort in and and even though you know metal music well I'm sure was poop. Goodbye my more established musicians, you know the the chops they were they were able to build where we're pretty incredible. I would say, you know, honestly that probably mostly you know what computers have done, have you? No more leveled, the playing field and you know I think I think probably Although I have my own stories of using computers to, you know, do things that I couldn't do myself on my own instruments. I think I think probably mostly, you know, computers. You know, have you know what more more made things you know, standard and, and probably, you know, more pop and more basic than You know, then then really inspiring


DEEP: people. I think, would you like, when you think they leveled the playing field, are you sort of speaking to something like a garage band, for example, that takes a recording studio out from being an incredibly expensive proposition back in, you know, even in the early 90s, but certainly before to being pretty cheap, now, to put together some good stuff is that is that kind of where you're going with that


JAY: you know, certainly PCS have democratized music production, Tremendously. I was just thinking more about your specific question about you know because a computer could do you know certain certain things. So much more easily than then a human was that was that inspiring you know humans to actually be able to play that way and I guess I generally say that that's not my impression. I remember using early midi programs and You know, going in the into the editor with a little pencil tool and being able to create a pattern that that I couldn't play myself and either slowing down the recording. So I could play the pattern like really slowly or just, you know, going in and hand editing a pattern and then and then playing it back and sort of being impressed with what I produced. But at the same time, sort of feeling a little alienated from it. Because it It to me as an experienced musician, I knew it was synthetic, you know, again that said I've heard many metal guitarists and metal drummers, you know, do acrobatics on their instruments that were, you know, pretty astonishing impressive. But I don't think that was probably inspired by computers.


DEEP: Interesting. So to say earlier in the, when we were chatting, you mentioned less that your, you see your kind of tech and music careers as being kind of, in sort of parallel universes if you will, or like kind of parallel tracks. Like, what did you mean by that?


JAY: Yeah, so so my these two worlds slammed into each other when I was in college. So, you know, just kind of continuing a little bit with my story from high school and my determination to be a professional musician. Mission. I went to Berklee College of Music for my freshman year in college. Which for those who don't know, is a professional music school, it's really known for jazz but but also, you know, pop music and session players and so forth, really talented people, go there and and, you know, learn the trade if you will learn learn music. From a bit more of a vocational perspective. And so you know, for example, my professors there where you know people who maybe maybe played one night with a well-known jazz musician, the next night they played in The Wedding Band the day after that, they were teaching giving high school kids or grammar. School kids private


DEEP: lessons. Yeah, I think I was on the other side of that a little bit. I when I lived in Cambridge across the river from Boston, a, I remember there were a couple of clubs where the Berkeley The kids would show up and you'd see the same set of folks, one night playing you know rockabilly and but all the way into the costumes and like but like they look like that was their thing. And then the next night you'd see him playing like a totally different genre like out in the Jazz group and then the day after they might be, you know, in like Marilyn Manson was big. She had some some of those like heavy goth rocker guys one night. So I always thought that always cracks me up because, you know, you just kind of growing up in the Northwest. That was pretty much like the antithesis of Of the, you know, the quote grunge movement. So I remember thinking like, yeah, but I also remember being incredibly impressed with these guys ability to not just pull off the music very like impeccably, but to pull off the personas and everything, like they just look like they came from that world.


JAY: Well, and what was taught there was versatility, because versatility pays the bills. And so it was a great experience. I only stayed at Berkeley for one year. It was a great experience in that it taught me that being a professional musician. Making your living with your instrument is really hard and there's a lot of really good people out there and, and, and you have to do whatever you have to do to pay the bills. And I was feeling more altruistic about music still at the time. And that really didn't settle with me very well. And so, so I. So I left Berkeley and I went first to the University of Chicago for about 18 months and then finally settled at the University of Iowa, but but all along and when I was, you know, went to these State schools, you know, I was, you know, very, very interested in just kind of broadening my horizons and I was starting to take lots of social science classes and ended up switching my major to anthropology, but I was still, you know, very interested in music. Zurich and music production. And so, you know, on the side I was, you know, using tuition money. And I was working when I was in school and, and paying a lot of my own bills. And so, you know, any money I could scrape together, you know, to buy gear. And, and one of the pieces of equipment that took me awhile to get ahold of, but I finally did was in the late 80s and early 90s. There were these products on the market for Doing multitrack analog recording on cassette tapes. And so you could buy one of these little tabletop units from Tascam and and other folks, they are about 500 bucks, therefore tracks, and they're pretty shitty. And so, so I come up with these compositions and I record the bass, and the drums and guitar and vocals, and they were frustrating because, you know, you Good work all day on a composition and then, you know, if the tape started to wear out or, you know, there was a, you know, an electrical burst on the power system or something, you know, it could ruin your convert composition and, you know, just in general, Fidelity was really low because the tape with was was so thin. I was talking to a friend at school one day and he said, you know, if you go down to the computer center at the school, See if Iowa said, if you go down to the computer center at the school, they'll lend you, the money to buy a computer. And he said, then once you do that, you can do me decomposition on your computer and it's much better than, you know, trying to do stuff on tape. And so I heard a bunch of things in that statement that really kind of changed my life. First of all, someone would lend me money. I mean that was just a crazy idea. In general, I think I had already ruined my credit at that point by you know that was the early days of credit card companies giving out


DEEP: credit a mountain. And yeah,


JAY: and thick. And then realizing like, oh, I have to pay this bill every month. But then, you know, I went and figured out what midi was and and really, you know, maybe in a lot of ways was my introduction to Computing, you know, given the given the fact that it's a digital medium, it's a 7-bit protocol. And,


DEEP: you know, for some reason I was just kind of


JAY: fascinated with with it so, so my next piece of gear. So I went down to the computer center, the guy said you can buy a PC and he showed me a PC clone running Das. He said that seven hundred bucks. You said or you can buy this other thing over here. This Macintosh, it's 2,500 bucks but it has a graphical user interface and it'll be much easier for you to learn. And again, you got to remember, you know, I wasn't a math head in high school, my friends. Our, but I was probably more friends with them because they were punkers to. And so I had friends with PCS like a really good friend of mine, from high school. My best friend had an Atari and he was he was always on and always doing what seemed like cool stuff with it. But you know, I never entirely, you know, understood. So I was intrigued with the idea of a personal computer, but, you know, now I now, I have one then, I Realized I needed a midi interface for my new Macintosh and I needed a multi timbral synthesizer to be able to do the midi equivalent of multitrack recording. But you know once I scammed all those kind of pieces and couple them together you know it was it was game-changing for me, the kinds of you know compositions I create in this is where we're getting back into what we were talking about before was was Astonishing. So so I got a copy illegally of a program called Master Tracks which was a multitrack midi sequencing program. And now I could record a measure worth of a Groove and cut copy and paste it or I could record a couple of measures of a Groove and go in and edit out the mistakes and then cut copy and paste those and and within moments, you know, had this beautiful sounding Being, you know, composition and I always kind of tended more towards analog type synthesizer sounds versus synthetic type synthesizer sounds. So you know, I would choose drums, that drum sounds and sounded more like real drums and and, you know, etcetera, Etc. And so that was really profound for me that power. And of course, you know, I was also writing my papers on this Macintosh. I was, you know, drawing things with Ain't, you know, I was buying all kinds of programs are, you know copying all kinds of programs from friends and and it was it to me it was the ultimate instrument


DEEP: the Mac itself. Yeah. All right. So with that I'm this is an AI podcast. I'm going to change channels and then we're going to try to stitch together. It's going to feel a little bit disjunct. But have you seen open a eyes new jukebox program? And


JAY: I have yeah,


DEEP: cool. So let's let's jump because I know that. So we'll jump into AI world. Then let's try to sort of stitch stitch these two worlds together a little bit. But you have any thoughts on these kind of generative kind of pre-trained Transformer like Technologies. So specifically you know like I just thought it was kind of a wild idea. Like what you know, we're both pretty and I think a number of our listeners are pretty familiar with this idea, of how you train up these things in the text world. Like, you know you basically train these Terms to try to predict like the next word or sequence of words based on, you know, this gigantic corpora of web documents. But the first time I saw with the Jukebox guys were doing kind of reducing all of a music composition down into this like lower, you know, heavily downsampled space, you know, I think 350 or so hurts and then within they're just trying to take the same approach to just simply predict the next samples in the model. Yeah. All right, I found it funky and just kind of wild and it seemed it just, but then, you know, you see the outputs of some of the stuff I think it was rolling stone. Did a, you know, an article and they taken a bunch of Nirvana, a thing kind of tuned on some Nirvana stuff and like was finishing a whole composition. That, that, you know, Nirvana. Never of course, made. But this thing authors, like, what are your thoughts on that whole approach? And it's, yeah,


JAY: yeah. I mean, it's I think it's time for me, it's very much of a natural evolution of where things have been evolving from for 30 years. So, you know, I was talking about midi earlier and, and, and just how democratizing the personal computer and midi wires, you know, remember the days to in the as late as the early 90s where, you know, going into a recording studio and cutting an album. Um could have been, you know, half a million dollar Endeavor or more depending on, you know, what you were doing even just to buy the two-inch tape, you know, that most major artists were recording their albums on, will those, you know, were role of that stuff was 500 bucks. So, you know, the computer and Technology democratized, you know, access to music production. And, and again, it started with midi, but then you mentioned GarageBand earlier, Which which, you know I also think was a really key, you know, addition to the Arsenal where GarageBand wasn't just midi, but it was also audio samples and Loops. Yep. And and you know, certainly, you know, it doesn't take a rocket science to listen to most pop music and and realize that it's composed of you just, you know, repeating Loops just, you know, hopefully arranged and you know, interesting and creative ways. And so So, you know, as of 10 or 15 years ago, you know, you could, you could fire up Garage Band, grab some pre-recorded Loops from its Library, you know, and put together a pretty nifty composition, just with drag-and-drop abilities. So you didn't even really have good, you know, to create a to create a really cool composition, you know, assuming you had a bit of an ear, you know, for these things. So so To make this leap towards, you know, these generative models, which, you know, can take some of these Loops or base audio signals and twist them in different ways. You know, I think is a natural evolution of where things have been headed. The big question is, is this good, or is this bad? So


DEEP: before we get to that question, which I think is a good one. So you think that if I'm reading you, right? You're saying the future of look. So jukebox is not doing kind of like track separated prediction, right? It's doing the whole thing, the guitar is the base that it locals. It's all swirled together just predicting the next sample points. So are you suggesting like the future? There is like kind of track separated instruments that are separated and maybe predictive from a compositional vantage like maybe you know, authoring scores or putting putting it down in the score and and kind of then leveraging. Within a garage band like context to try to, you know, add in to add an instrument to render or something. Like, how are you thinking about that like Bridge those two worlds for me a little bit?


JAY: Yeah, I mean, I think you could go in a number of different ways, you know, think about the style transfer stuff that we've seen, you know, in on the visual side with, with convolutional neural Nets, you know, they're you're taking the higher frequency components of One image and hanging them on the lower frequency, bones of another image. And so, you know, you can certainly easily imagine that Paradigm, working in music, it's a little hard to imagine like how things would actually sound. But you know, you might say like, hey I really like, you know, this this Rihanna song and I really like the music production values on it and I really liked it. Homes that are on it, but put take all of that and put it on top of this other chord structure. So, you know, we know of course in rock music and in pop music, that that chord progressions are fairly standard 1/4 + 5 is incredibly standard, you know what 162 is you know another something that's really standard so you know those I think could be the bones. You know, that lie, you lay the higher frequency elements of how the song sounds, you know, onto onto those bones. So, that's, that's one way that's just the, I want a new song and I wanted to be a mix of things that sound like this and things that sound like that. And then but I also, I think, to your point, you know, working with individual tracks is pretty interesting to, you know, including Track separation, which to me, is sort of a Holy Grail, you know, if you can there's a bunch of multitrack stuff floating around on the web right now and I don't know where the how this stuff got out or where it came from but you know there's a there's a there's a guy whose videos I watch quite a bit. Rick beato, who's a, who's a musician and producer, whose about my age and is a lot of really interesting content out there. He Finds these multitrack recordings from the 70s and 80s and 90s and so forth. And then does song analysis. And when you're able to isolate those tracks, you know, that's when you really have the most flexibility


DEEP: because now you can put all kinds of effects on the guitar on the bass, or you know, whatever you


JAY: can just change it out. I mean what if you took a Led Zeppelin song and you know, pulled out the Bass track and you added your own


DEEP: that's interesting


JAY: or Are you know or some, you know, the vocal tract or whatever. So you know, again there's there's precedent for some of these things that really go back up surprising number of years. If you think about it now. And in this case, it started with sampling, you know, the rap artist in the in the 80s and 90s were were sampling, you know, chunks of, you know, recognizable pieces from well-known tracks, some of them got the pantsuit off them for it, but but It's really kind of an extension of that. I'm going to come up with this whole new song. You Know based upon this really well-known Baseline and I've got the flexibility and freedom to lift that Baseline you know right out of this right out of this recording and now it's a canvas and I can, you know, do whatever I want with it. I mean, that's that's really powerful.


DEEP: Yeah, I think I mean there's these this this is kind of The Crux of it is, what is left of the musician in this world. And is it good? Or is it bad? I mean, you know, if you rewind to the 70s, most musicians had their instrument had their group of folks. They played with maybe they swapped instruments. They had a good time. Maybe they wrote it down, you know, but like even then that might be a big step for those guys and that might be quite high tech, you know? And, you know, to be a composer in the, you know, during like Mozart Sarah for example, was like a whole other thing. They had to hear it in their own head basically and really, you know, be quite intimate with the, you know, with the notation and be able to write and then they had to have access to it, you know, the ability to render like a whole orchestral piece or something. All of that stuff going back to your democratization point. You know has it's like anyone can grab a garage band and be a legit company you know, composed stuff now and then swap in other instruments and stuff. It feels like it's a great thing but most musicians I know at the end of the day love just playing music together live and communicating, you know, via their instruments and you know their settings and then they almost like go into a different mental mode like more like the software program. More kind of mindset when there are mucking around in electronic world. Like, where is all this? Yeah, like is this good? Or is this bad? Like what in the, you know, especially if you start pulling in the, the toolings on its evolutionary path. But now we start talking about, you know, machine learning AI techniques. You know, track separation is one thing but downright straightforward authoring of drum beats of Loops of, you know, all this like yeah what what is it? Is it good? Or is it bad? And


JAY: yeah, I mean I think it's probably in the eye of the beholder. There. I think to some, it's, um, it's good are great, you know, to others. It's, it's really bad. And, you know, there are these sort of two elements of creating professional music. There's the compositional part, but then there's the production part. And so, some, some of the, some of your listeners may know that, you know, in fact on most of their favorite recordings by their favorite artists, the role of the producer, you know, is is a really big role. Role in a silent role to the consumer of the music, but really the person kind of running the show and the recording studio. Historically, the producers are really empowered by this technology and, and you know, many folks, especially, you know, I'll just, you know, call myself out as being an old fart, you know, many many, you know, folks from from my generation. You know, complain that most popular music today is a producer and their laptop and a singer, someone with a voice and an auto-tune program. And so, you know the the the the you know, the the feeling I think is is that, you know this this is an art and and it and it's devoid of talent as well. I suppose. Really think the, the auto-tune feature, you know, which is surely leveraging, state-of-the-art AI technology. You know, is playing a big role there, and so, I think there's a lot of I think there's a lot of folks who are not aware of any of these Dynamics and they just, you know, turn on their music and listen to it and enjoy it. I think there's a lot of people who long for the days when the people producing pop music were Talented and can play. I watch a lot of videos on YouTube from 70s, the 70s and 80s. Especially when the only way to see your favorite bands, perform was on TV late night whether in music videos or you know, live like on Don kirshner's, Rock concert was the late night TV show that used to have popular bands on and and you know that stuff's all on YouTube now and when I go down and read the comments, it's the same comment over and over again again from probably somebody from my generation. Saying, man, I really would I long for the days when my favorite artists could actually play because you had to you


DEEP: had, I mean, I feel like, you know, music, like art. I can kind of use this example a little bit in the art world. But, you know, they, a lot of times folks, you know, used to ask. Like, what, what's, what's actually the role of? Artists and you know, artists are pretty amazing at making lbs and they're you know, even back in the 80s we everybody, you know, as soon as everyone the mass is found that Milli Vanilli was lip-syncing. I mean they were pooh-poohed. So and a lot of these singers who are maybe auto-tuned in the recording, studio can still sing, you know, like the so they play live and yeah, they might ought to tune them here and there but they, you know, they still can sing, but there's a lot of folks that, you know, that, you know really Go more spare and get these build, these organic audiences, like Rodrigo and Gabriela come to mind, you know, a pair from Mexico City that play like, you know, pretty heavy metal. But like Flamenco style, largely pretty spare, the musicality and the, and the rendering ability is, you know, I'd say would even Trump the, you know, the, the compositional aspects, for those guys. Yeah. So yeah, I mean, I think, It feels like the every time something gets automated it's no longer. Okay? To be the artist who does the thing before it got automated mostly right? Like when photography it evolved and we started getting photography all over the place, it was no longer. Okay? To call yourself a pain or simply because you painted realistic imagery. If you are going to still dabble in realism, you had to be doing something different with the realism that can't just be captured by the lens. So it feels to be like The Optimist in me wants to say that to the extent, that things, get automated that musicians are very creative people. They will bore a quickly of anything that smells like simple-minded automation.


JAY: Yeah, I think he's real musicians will for sure. There's go course other entities in the music business, you know, there's there's again, there's producers, there's lawyers, there's record companies and a lot of those Dynamics have, you know, greatly changed over the years and yet, you know, there's still the same, you know, to some degree as well. I mean, I do think you're seeing more artists now that that our producers in their artists as well and and so they're composing and they're producing and they're writing and creating. So I think that that that role kind of has has been more more elevated, but then there's a lot of people out there too that just want to play just to play and I try to do both. So you know, when I'm doing composition, In GarageBand, by the way, which I, which I love because it's simple enough that it doesn't get in my way creatively when I'm, when I'm, when I'm in that mode and and a lot of other software I used over the years was probably more than I needed. And, you know, created problems but you know, I will, you know, at a convenience and efficiency, I'll lay down two bars of something and then quantize it and then You randomize it and then copy and paste it. And, and that gets a composition together really quickly and it's proficient. Maybe I feel a little bit guilty if I play it for someone and they're like, wow, did you play this said? I'm like, well yeah, kind of but then I, you know, I also try to jam with friends and,


DEEP: and, and we actually


JAY: play and so I, I hope that the playing part Makes makes a comeback and and learning guitar or learning some instrument as a teenager, you know, is, you know, returns to being a rite of passage.


DEEP: So I want to talk a little bit about this. This kind of new world that opens up when you can leverage Ai and it's different from what we've pretty much been talking about. So far like singer-songwriter artists bands playing in front of, you know, No, people are recording to for the sale of, you know, records and stuff. But there's this new thing that's kind of being enabled. I would argue by a massive demand in low-cost music generation and computation. So if you think about video games, you've got to create all of this, you know, audio to go with the video games, at somebody's moving through the game Space and so, you know, there's a company like, you know, a IV, a Ava You know, has this kind of dynamic, totally personally tailored to the scene in the game, kind of music generation. And then there's other companies that are doing things, similar things, like you like composing on the Fly by weather, or by emotions. And then you've got, you know, there, I feel like these companies are kind of trying to address the sort of dearth of easily call it writes accessible music, you know, for all the content being generated out there. So like, you know, am Isaac is sort of an example of that. So what is that like when you when you sort of put the traditional music world off to the side and you put on your your Tech kind of future hat and you start seeing a you know, everything from Smart Homes and you walk in and you know because at the end of the day the actual known public music catalog is like I think there's like 10 million songs published totally. That are like kind of column record label equivalent stamp and of those maybe 50,000 or Kind of get played over and over again and probably, you know, a thousand of those that drive us nuts. So but there's this whole other world that could emerge where you could start to like have very personally tailored or environment, Taylor music, what are your thoughts on that? And what role a I might play in that?


JAY: Well, I think like a lot of things like news stories and other kinds of content, you know, music has already gotten caught up in the recommendation engine collaborative. Filtering Bubbles and, and, you know, so it's really a poster child, you know, for that. And again, is that a good thing or a bad thing? Well, you know, it's great when I have Pandora or Spotify on and I am wanting to play some Reggae with when I have some friends over and, you know, we're enjoying music and having a few drinks and so forth. But the course the bad part about it is that I'm not being exposed. The new music, you know the experimentation over-exploitation formula you know gets gets gets out of whack and so you know there may be a great song out there. You know that I would really enjoy that. I never get a chance to listen to because you know the recommendation algorithm says well we don't think he's going to like it. So you know I think that Pandora's boxes Has is already open and you know, whether it's a good thing or a bad thing. I think, like the a lot of the other things we've talked about, it really depends. The other thing I would say is that, you know, I've always wanted and been curious about Dynamic music. And so, you know, we know, we know a given song has a structure to it and sometimes there's alternate mixes of songs like an extended dance mix. Or something like that. That is a special purpose. But you know, song is like, you know, you're going to start with A Chorus and then you're going to or you can start with a verse you're going to go to a chorus, then you're going to go to a bridge and then back to the chorus and you know, blah blah blah. But I've always sort of wondered like you know what if what if that was randomized and it was back when I've got my first CD player in the early 90s that I wondered about if we start to see stuff like this and maybe even Even like, you know, it can learn a preference so that you know you might like to hear a particular song, being intro Verse Chorus, Verse Chorus Bridge you know tag outro but I would want to hear that same song with a slightly different structure. Maybe maybe the bridge come sooner. You know, I think things like that or maybe I would want the mix to be different. Then you would, and it would could dynamically adjust. And so, you know, that that could happen again. If you could pull tracks apart on the fly or if you had the individual tracks stored together but separately, you know, that that kind of personalization, you know, maybe a frontier II, don't know. I mean I think it's one thing is if the technology can support it, it's the other thing. If whether people can Care.


DEEP: Yeah, I mean it feels like the the personalization lens that you're talking about. It feels like I can play in with other other kind of emerging forces, if you will. So like one of the things that's always driven me, nuts and mute is music. You know, you get in the elevator Heat this, you hear this kind of garbage, music stuff and then it feels like, you know, you got a smart elevator. I'm the only one in the elevator. Why the heck are you playing music? Seems like an inevitability at some point but all of that's kind of you know like like track selection and you're kind of getting at like what? Like maybe a helping people think about what is underneath the surface that they really like from a structural Advantage from like I love music with like really dramatic you know you know sound and like sorry volume differences, you know a lot of dynamics that way. Whether it's you know classical or you know or rock or heavy music that that to me is like you know really important. All right I'm sure we could talk about this forever. This has been an awesome conversation. Thanks a ton j4 for helping us just you know scratch the surface here.


JAY: Awesome. Yeah, thanks for having me, always fun to talk about this


DEEP: stuff. Cool. And I look I look forward to jamming together so that's all for this week folks. As always, thank you too. Listeners for tuning in. If you've got a chance you know we're a new podcast. We'd love you to jump out and give us a rating. If you like what you're hearing, we've also got an article that's dropped onto our website. That's all about this music composition topic that we've been talking about here. Today, we go into a lot more depth there. So if you can find that xyonix.com/articles that's x-y-o-n-i-x.com/articles, thank you and over and out.