Unleashing Creativity: How AI is Transforming the Music Landscape with Staccato
00:04.57
mike_flywheel
What's up everybody. It's Mike we're back with the pitch please podcast and today I've got Jeff from stacato they're an ai music co-writer that helps everyone unlock their full creative potential as a creator I'm excited to learn more about Jeff about this new Ai innovation that they're working on. And a bit about how staccato is going to change the world of of music and creators welcome to the show. Jeff maybe start with a quick introduction about yourself.
00:28.65
Jeff
Absolutely thanks for for having me Mike. So yeah, my name is Jeff and basically staccato is a platform of Generative Ai tools for music and lyrics. So essentially we help songwriters and music creators overcome any writers block or just Spark New creative ideas from scratch. It's kind of like a music library or database at your fingertips so you can always stay in an uninterrupted state of creative flow and then the most integral part that we always want to get out. There is that we never want to replace the human creator and instead we're only looking to enhance our abilities in the way they create as all this Ai technology unfolds. Kind of so rapidly nowadays. That's our biggest thing that we always want people to know.
01:08.12
mike_flywheel
Super cool. Let's let's dive in and a little bit about your background. So right now you're the founder or co-founder of of staccato co-founder of Staccato um, and obviously you know you've been doing this for for roughly how long.
01:15.67
Jeff
Cofounder Cofounder yeah.
01:23.29
Jeff
Ah, so I started all of this back in my ph d so probably four or five years ago is when I was just kind of dipping my toes in this sort of generative Ai space. But in terms of Ai and music probably even longer maybe in 7 years at the start of my pc.
01:39.19
mike_flywheel
Oh Wow So do was this like ah a birthing of your Ph D like where you kind of just this is sort of your major thing after your Ph D or tell us a little bit about that or what do you even did your Ph D And maybe that's going to be super relevant to this.
01:51.49
Jeff
Sure. Yeah, so I'll even go back farther than that. So I I basically like I started playing guitar when I was 13 and then a few years later I was like completely obsessed and then I started playing professionally on guitar and keyboards for anybody that would hire me a lot of music theater.
01:58.26
mike_flywheel
Okay.
02:09.88
Jeff
So There was some point in high school. Definitely where I was like I just have to my my life has to revolve around music I don't care how it just has to have music in some way shape or form but I was always really interested in composing at the same time and music theory and all just kind of how music works So when I went to Western University I took. Music composition and music Theory. It was kind of a combined thing back then um and sort of I went the whole underground route and you kind of do the interesting stuff that you can, but it was grad school really where they let you do kind of whatever you want so long as you can back it up So I was always interested in how I could apply tech to Music. So My early projects were kind of it was like coding it was ah an excuse to like teach myself to code. But then say that I was still working towards my my masters and so nobody was getting annoyed. Um, and so I was doing live coding and music. So Basically your coding functions that will then make the music and it's sort of like what Djs do.
03:07.51
Jeff
Um, where you're tapping along at a at a computer and then the music comes out so I had fun with that and I was trying to do things where the audience could collaborate with what was going on stage again with sort of coding little web apps where the audience could sort of get into this. Web app and then as it performs was going. They could hit little buttons and they could make music too or their phones were kind of surround sound extra surround sound for for what was going on stage so I was already like thinking about how to create web apps but with not any particular goals of being in business or starting a company. It was just purely out of. Interest and how could I make noise and music and in in new and creative ways and then when I kind of kept going with my ph d I thought I was still looking for what what would be kind of interesting to do and this is when Ai was starting to have another resurgence and as soon as I saw gp 2 which was the kind of precursor to chat gbt. That's when I was like oh this is it. This is the one so I was working on how could I basically rewrite the the models that work for ggptwo back when it was open source into something that would work for music so instead of just looking at the text and predicting what words should come next. It was looking backwards. But looking up and looking down see all the chords and harmonies that were going on and then could it predict what was going to come next. So that's what I did for my my ph d and then when I left school I just wanted to keep doing it so that's where the the company sort of formed was just out of grad school.
04:37.59
mike_flywheel
That's super cool. There's a lot of things I want to dig into there but at its core you've sort of been a music buff like your whole life and you've sort of been persistent in that you are going to forcefully find a way to make your passion for music.
04:40.51
Jeff
Um, for sure honor flag. Yeah.
04:51.10
mike_flywheel
Your career path and so you kind of pursued that in school we're working on some things alongside but where did this like tech curiosity come into this like I can't imagine everyone just picks up randomly without even going to school for it and decides they want to be a coder. Is. It. Like spawned out of your desire to create and be curious or is there something else that sort of inspired you to start playing around in in the tech space and starting to bring these things together.
05:16.43
Jeff
Yeah I think for sure. Um, there's just something if you like music composition I think there's something inside of you that is happy to create in kind of any way shape or form. So I like other hobbies I like like cooking and photography are 100% distill that kind of creative. How can I make. Something from some type of ingredients. So the ingredients could be ah music notes music chords stuff like that. It could be food. It could be what you see outside but they're just or if for programming it could be functions and variables and stuff like that. So. It's just. Wherever there's some sort of ingredient I think I find joy and how how can I put them together into something that's kind of interesting. So I think there was definitely something there that it's just such a natural fit between music composition and coding I mean there there are studies into. Math and music and being some relations there. But I think for me, it was just sort of yeah you got this blank page and how can you do something kind of interesting with it.
06:16.33
mike_flywheel
That's that's super cool and I'm glad those I Love how you kind of found the common thread across all of your hobbies there being the the mix of ingredients. Um, now entrepreneurship was that ever a consideration or like before you jumped all in here.
06:20.33
Jeff
You have exactly.
06:33.30
mike_flywheel
Um, with stacotto like have you always thought that maybe entrepreneurship would be something or maybe it would just be like doing your own thing but in in music um was that like something that you thought of earlier.
06:42.13
Jeff
Yeah, it was ah no, it kind of just happened. Um, so I kind of always figured the route I would take is um, just stay in school and then try and become a professor after and teach. After that. That's what both my parents do. So I think I just saw kind of the path there and I had a lot of help too of how that could potentially happen by doing certain things. But then when I finished school I just needed a little break from theorizing and writing. And I just kind of wanted to create a little bit so I just started consulting and just sort of and because this was when Gp 3 was taking off so everybody was looking for ways to apply g P three into their business. So I was just consulting in terms of here are my skill sets that I've done over the past. While and most of it had to do with language and GPThree so I had to understand that before I could make it work with music so I was just sort of helping other people there and kind of learning what it means to run a business on a very small scale like just for yourself and I had some help along the way with different people. That sort of I was around and sort of how to make things better and I think as a professional musician you're you're self-employed and you're kind of already learning to be entrepreneurial just from that you learn how to go to a gig talk money sell yourself, um, sort of.
08:09.23
Jeff
Find value in what you can provide and how can you sell that and can you kind of bring more value to the table and then thus make more money for yourself and can you live off of that and when it's an extremely competitive and difficult world. So I think there was just little pieces that were kind of falling into place but it was never a big desire to be. An entrepreneur. It just sort of kind of happened and then because I wanted to make staccato happen. So badly. The the rest was just sort of find whoever could help me realize this goal and the the biggest person for that is my cofounder Jason who sort of went to iv's. Business guys and accountants so he brings all ah all that stuff the table and I can just pick his brain constantly for all the the missing stuff.
08:53.50
mike_flywheel
That's cool and I love how you're kind of describing themes of things in your life that are so related because I think sometimes people think they need to be like cut from this entrepreneurial family cloth. Or they need to have been a 5 ive-time founder to start something but you've talked about a lot of amazing core attributes and skillsets of other things you were doing whether it being an independent artist whether it being a creative and creator of you know, great food, great photography, great music but also great. Great applications.
09:21.33
Jeff
You.
09:23.57
mike_flywheel
Um, now you you touched on a piece there which is like you just felt like you had to start Stacato What was that moment like what what kind of pushed you over the edge and I want to even like go back after and talk about Gbt two because most people probably don't even know what that is but let's just talk about like what pushed you over the edge on stacato.
09:39.97
Jeff
I think it was happening during grad school when I was building it for for school I mean it was much more of a classical music. Um kind of discipline that I was in so everything was really revolved around how can this work for writing a piece of music. Because I had to write a string quartet as part of my dissertation and so part of the goal of the what essentially was the Mvp of staccato I used it to write the string quartet so I would pass ideas back and forth with it and it would inspire me and so on so forth until I had a 20 minute piece but I always felt like this would work really well for pop music especially in the sense of I wanted to build a lyrics application to it because I always sucked at lyrics I would write guitar parts and everything like that. But then I would it would come to lyrics and it would just be. Looking around a room and trying to figure out whatever was right in front of me. There was no poetic aspect to it. So I thought a lyrics application too would be great given all these text tools that are out there so I was already thinking I want to bring that. But I think that because there was some. Stifling is a hard harsh word. It's not that they were doing that they're just trying to keep you focused so you finish your your work but I always felt like it needed to be something bigger. So I think as soon as school was done I wanted to sort of um, a bring all that stuff that I was thinking about already to the front and then b.
11:08.61
Jeff
Make sure that I was making the decisions and I wasn't influenced by anything around me so that kind of naturally meant starting a company where okay, it's your thing and then from there you can make all the decisions whether they're right or wrong or focused or unfocused at least you get to to make those decisions. So I think it was sort of it was growing and Then. Pushed me over the edge I think was doing the consulting and you're working on all these other people's projects some are cool, some not as much and then at some point you just go I don't want to work on anybody else's stuff anymore I Just want to do music like I mentioned kind of earlier and get back to that because that was always the ultimate goal.
11:44.63
mike_flywheel
Super cool now um, trying to figure out how there's so many things I want to talk about this is super interesting. Maybe let's dive into a little bit around this this space I guess actually hold on. Well.
11:48.49
Jeff
I.
11:58.74
mike_flywheel
I don't want to even jump ahead because I still wanted to redo the pitch and and talk through a little bit of this. Maybe let's talk about like gp 2 now I think like everyone probably like kleenex level knows what chat Gpt is but you're talking about this thing called GpT 2 when when was that what got you hooked.
12:00.82
Jeff
Yeah.
12:11.80
Jeff
Um, sir.
12:18.18
mike_flywheel
Or interested in in Ai because Ai well talk about when gp 2 sort of inspired you. But I'm curious. What got you curious and Ai at that time because that wasn't like in the last year and probably most people have heard of or really got curious. Oh ai in I don't know last twelve twenty four months maybe
12:34.31
Jeff
Yeah, yeah, so I think it was around 2017 is where Gpd two came out and I think what got me hooked was that they posted something and there was a gp one. There was the first one gp I should say stands for.
12:36.39
mike_flywheel
But before that it's been around for a while. So what guy you hooked? yeah.
12:54.35
Jeff
Um, ah, generative pre-trained model. So essentially what it's saying is that you can use this thing that they've already pre-trained on it. It can speak english essentially and you can do things to it to teach it little other things like maybe teach it something specific on how to write an academic paper. And you don't have to start from scratch and teach it english it already knows kind of maybe let's say a high school level amount of english and then you teach it a little bit more so that's where the the pre-trained aspect comes from and then what happened with the move from Gpd one to gptwo was they. Didn't release the best model that they had so there was these smaller ones that were pretty good but the best 1 they said it's too dangerous to release. We're scared about what this could bring I think I saw that and just thought oh that sounds like the coolest thing ever. So I need to know. All these other models up to that point and they did obviously eventually release it and the world didn't end so that's good and over now at gpfour and the world still hasn't ended so that's great, but there was something about this idea that it had learned english in their minds. So or I think it was mostly english at that point it might have had some other applications other languages. Not as good as it is as translating nowadays. But there was something about this idea that they had trained it on english text to such a degree that they were now worried that it might sort of get better at writing text than humans or be on par with humans that I think I was just sort of like oh I've always like tech.
14:24.26
Jeff
But is there is there actually that ability that you can train a computer on this skill that is your own skill that you've kind of spent your life on to such a level that maybe it rivals what you can do so I was partly seeing if that was possible but partly just understanding. What. Is behind that that might bring it to that level and is it is that even possible and I think the more I spent researching and doing it the more I sort of squash those doomsday sort of opinions of it and it's really not competitive with humans. You you can train it up to some level but the humans are always 1 step ahead and I mean it remains to be seen what happens in the future. But at least at this point it just isn't as good as what we can do but there was something interesting. There. Okay, if you can train it up to this point. What can the human bring to it and can it enhance what the human can do. And does the human always stay ahead of it if it's always enhancing what the human can do. So I think that's where the kind of the interest came from just seeing the initial sort of Gpd 2 papers that I was reading? Yeah yeah, so ah.
15:23.25
mike_flywheel
That's cool and so are you the technical cofounder of of stacato so you self learned because you didn't go to school officially for this, you self learned coding.
15:33.60
Jeff
For yeah.
15:35.79
mike_flywheel
The applications and implementations of Ai just out of your own curiosity.
15:40.45
Jeff
Yeah, so I yeah I was in music school the whole time that I I did this so I wasn't in compsi I was lucky that one of my advisors throw my ph d um, he did have a and I sought him out like I tried to find somebody that could help me in that. Um, regard and but he had a big background computational linguistics and kind of history of computers and all that kind of stuff and he was a good coder himself. So he taught me a lot but some of it was just buying books off Amazon and working through each example and and problem and then ultimately. Once it felt like I had gotten to a certain level. It was just downloading all the papers because back then everything was open source but now everybody's trying to monetize it so it's not always open source. So I would just compare what I was learning to what the cool things were and because I had a goal I was pretty driven towards it I think if you just start compsite from scratch and you have no goal. And you just start at day one variables or something like that. You don't have a goal so it's not as exciting. But if you know what you want to achieve with it. It's it's sort of easier and then like I said it was in grad school. So I just built it into my project so there was always something that I was working towards and it just sort of sped it up from there but I was pretty driven to happens.
16:49.95
mike_flywheel
There was but it was like the tangibility of what you were working on that you are sort of like I'm not solving ah a hypothetical challenge like I'm actually trying to get this to do something for me so you're yourself learning it alongside your Ph D which is super cool. Um you.
16:55.59
Jeff
Yeah. Yeah, yeah, and it wasn't like I and it was like I'm gonna have it write music in a month. It was like maybe next week I could have it pick a note out of the 12 notes. What would be the highest probability note to pick and is that the correct note to pick so it was very so. Slow. It wasn't like it happened sort of over one semester. It took quite a while but and there was just little pieces of the puzzle I had to figure out along the way.
17:24.88
mike_flywheel
Um, so what was like the hardest part of that. Would you say.
17:30.30
Jeff
Um, that one was not too bad because I had time. Ah and when you're in grad school. It's you're just sort of in school and you know you've got these ultimate deadlines and if you're really passionate about what you're doing. It doesn't feel too. Bad or too hard at any particular time like you just have all this time to work on it. What's hard is I built that for myself and only myself but when I came out of school and I wanted other people to use it that was hard was figuring out. Okay, my coding skills are not where they should be for making a server and letting tons of people. Talk to the server at the same time. So the hardest thing was I think was scaling it so that even just more than 1 person can use it and the thing was like I probably had some hackway that worked but probably wasn't the optimal to make it work so the hardest part was definitely how can I make it so it's ah. More than just me can use. It.
18:28.75
mike_flywheel
Yeah, so making the the thing that you made for a singular purpose scalable to to help others. Um now in this ai space. Ah you have to obviously be pretty engaged with.
18:32.32
Jeff
Exactly.
18:42.30
mike_flywheel
You know the pulse of what's going on what are some of the most interesting things you are seeing and are some of those things making their way into how you're thinking about stacato and then soon we'll talk about Stacato deeply soon. But I want to just get your take on like this ai space some of the most interesting things you're seeing some of maybe the scariest you know you said. There's all these like Doomsday predictions back then obviously that that's not there. But there's some guardrails that we probably have to be aware of like what are you kind of kind of seeing.
19:08.65
Jeff
So Um, I don't pay as much attention anymore to image and video creation as I used to just to stay kind of focused. Um, it's all but I'll circle back to that too from what I hope doesn't happen in that kind of World. So but I do pay a lot of attention still to what's going on in these text models. Um, and I'm always very interested in what openai is doing especially with chat gbt because the Gp 2 version that I was talking about all it could do was you would input some text. And it would just continue the text for you based on sort of what's the probability of the next word being the most important but now the fact that you can do translation you can have it draw tables for you. You can have it code. You can do all these different things. That's the most exciting one in terms of what we're doing is okay, we have. Something that can continue the music for you. But how how can we make it more exciting and how can we have it do other things or fix things that are in in the past of your music that might be a good to go and seek out something that seems a little bit off and then just add a little correction like it does with coding that's the thing I keep sort of. Figure on the pulse of the most is just sort of how these language models are sort of getting better and better and improving and just the interesting use cases for them. That's always kind of exciting for what could we? How could it sort of come over to music in some way shape or form or performers or anything to do.
20:34.66
Jeff
More than even just music creation. But the other aspects of of music. So I find that interesting but to go back to the image in video the the nerves that I would have around that aren't so much the ability of what Ai can do. But it's how people. Might use it because it's cheaper and it's um, good enough I worry about a good enough approach to things. So let's say for movies. What if you could generate a good enough movie for an hour so we already see that maybe movies aren't always as. Interesting as they used to be so it's more just like this movie 2 this movie 3 and 4 and and there's no sort of new content. It's just what we'll sell so what happens if the Ai can replicate um, what is marketable and what will get enough buts and seats that they make money. And so it just becomes okay, this isn't is 80% as good as a human movie but it will sell pretty much just as well. So let's go that easier route because it'll be cheaper and then ultimately we'll make more money and then it's the same for music too as I would worry that. Some of the audio related ones where we do we focus more on kind of the creation side of music. But once it's been recorded could record labels eventually have these tools that they can make songs personalized to people so you get your edge sheer and sounding song. But it's 100 % personalized to your life and your experiences.
21:49.85
mike_flywheel
Here.
22:05.10
Jeff
So are you more likely to buy that over the more general one Edge Sheeran does and then it's got this good enough approach where it's good enough to make you buy it or listen to it but it's not so good that it sort of changes music and keeps music moving. So I worry about this kind of static. Thing that might happen where music just sort of stops here and it's never really getting any better but it's just sort of as good as it's been trained on up to that point that that'd be the most worrying thing that I could see right now over the yeah.
22:35.60
mike_flywheel
It's an interesting concept. So like so in an ideal world. Probably what you empower creators or you'll love both to persist where there's the good enoughs but you still have original content and original pieces. Which push the category forward and the reality is you just have more content Now. Maybe.
22:49.70
Jeff
Exactly yeah, but the the thing too is that there there is research to suggest that when we listen to music we want to hear I can't remember the percentages. So don't quote me on them. But it's something like we want to hear 80% of something that we recognize. That's why popular music is so popular.
22:59.20
mike_flywheel
You had no words.
23:07.69
Jeff
So we want to hear drumbeats that sound familiar and chord progression sound familiar. But we want some little percentage of something that surprises us that we've never heard and that's what makes us go. This is my new favorite song so part of me thinks too that if there is this thing where Ai can. Regurgitate a lot of music or movies that is is good enough is that we'll still crave as humans that surprise that the ai just can't do and then that's where you still have the humans bringing in their creativity. So that's that's the the hope is that it will still just kind of will always be ahead of the ai in that sense.
23:41.87
mike_flywheel
That's that's super cool I Love like I'm like putting those pieces together and maybe envisioning how that world works but I'm still loving your your human-centric approach to this as an assistive tool and there's these still these very important categories where humans play a very important role.
23:51.41
Jeff
Any knew.
23:58.74
mike_flywheel
And think that's the biggest dialogue that people need to have more openly is like yeah there's some realities of where ai is going to you know, lean in but we have to look for those real opportunity areas I think I Really love your example of how humans crave some percentage of difference.
24:00.43
Jeff
Absolutely.
24:15.94
mike_flywheel
To make something their new favorite song and how Ai can take things it knows but it'll have a hard time creating these unique elements that that's probably where humans are really going to be able to excel. Um, in in this piece. So let's maybe bring that back to to Staccato um, actually but before we even dive into I'll get you to redo your pitch so can talk about the business but how did you come up with the name stack autoto.
24:38.87
Jeff
So there was a few things that I wanted it to not be 1 being just because of Ontario company I didn't want it to be the general Ontario and 7 numbers. Whatever it is like you see for your landlords and stuff like that. I didn't want it to be like that because it's boring and I didn't want um it to be a lot of companies and not to knock them. They have ah it'll be something Lee or something eo so I didn't want it to be like create eo or create Lee. Or musically or anything like that I wanted it to be something kind of different so I was going through all the various terms that you see when you're playing music or writing music. Most of them are italian terms and then I thought back to my holding company that I had started for sort of other projects that I was doing and for that one I had just. Had the same approach I didn't want it to be an ontario corporation with a bunch of numbers and I decide I would make it a music joke that at least made me laugh I don't know if it made anyone else laugh but it was um I called it tenudo which means to hold or to hold the note longer or held note and I just thought it was funny for a holding company. So I think when I was thinking of other names I thought well if if it's going to hold sort of a company that I'm working on. Maybe it should just be the opposite and staccato is essentially the opposite of tanuo where you have a detached way of playing the note rather than a held way of playing the note so that's just really where it came from and then my little sister who's ah in design.
26:09.10
Jeff
Went to school for design. Um, she was helping me create a logo and I think we just found some logos that looked cool and so the name just kind of stuck Once we had a cool look for it but it kind of worked out I still like it So it's definitely. There's no remorse over coming.
26:21.92
mike_flywheel
The the music nerd in you is coming out I love it and I love how it has like a real core rooting in like a music element that you know you've got a layer too deeper but it it has real meaning for you which is which is cool.
26:26.78
Jeff
He's elected.
26:36.67
Jeff
Yeah.
26:38.73
mike_flywheel
So maybe give us the pitch on what Staccato is again and then I want to talk a little bit about you know this category. Um, what else is in it and and some other elements of your business to help people understand what you're what you're building here.
26:49.27
Jeff
Yeah, so ah, essentially yeah, we're a platform of Generative ai tools for music and lyrics so songwriters composers music creators producers any of those types we want to help them out as much as we can and the biggest. Pain that we're trying to overcome and this is sort of what I built it for initially was writer's block. So for me I had to write a 20 minute piece and I'd never written anything that long I knew I was going to have writer's block if I had it for 4 minutes I'm going to have it for 20 minutes times 5 so I I tried to help. That standpoint of what if you never get stuck. So if you think back to university and excuse me in you're writing an essay some days. It just flows out of you and some days it takes three days before you can write another couple of words. What have you never had that moment of. No inspiration and there was always just some spark whether you use what we generate for you wholesale or it just gives you enough of an idea to keep moving. We just want to give you this music library database at your fingertips at all times. So. There's always a suggestion that can happen based on wherever whatever point your music is at or your lyrics are at. And then so because of that we think of it more as an ai co-writer and we're always just trying to enhance what you can do. We're never trying to make all the decisions for you so we try to avoid anything where you type in a genre in a beats per minute and it generates the entire music for you because there's no.
28:17.33
Jeff
Interest there and then you you haven't done Anything. You've just suggested something based on a couple Keywords We want you to be the creator and you have to be the creator and we'll just sort of help you along the way and keep you creating and that's the most important part that we see is that we're just sort of. Making sure that these musicians songwriters creators are still the most important person in the process of creating new music.
28:39.41
mike_flywheel
So steal a Microsoft Termin and branding here. You're the copilott of of music creators everywhere. That's that that's that's super cool now talk to me about this like category um of.
28:43.44
Jeff
Yeah, we like to think that we are yeah or the chat G Bt and music.
28:55.41
Jeff
28:57.62
mike_flywheel
Assistive music ai and music. Um, what else is out there. You don't have to drop your competitors' names and give them free publicity. But. Like what's the spectrum of what people are using to create music assist them and create music replace them and create music. What's like the spectrum of what's out there I'm I'm super not tuned into this industry.
29:17.42
Jeff
So if we look at just ai music. It's actually kind of interesting and because we're trying to differentiate ourselves because ai music can meet a lot of things. Um you could argue that Shazam is some level of ai music where you play at a song and it's using kind of machine learning algorithms to figure out what. That song is so there's still ai music there and then some of it is I want to play guitar to my favorite song but there's a guitar part in my way. So there's these ai music tools that will actually remove that guitar stem from the track so you can play yourself and feel like you're the main person in the band. So. 1 thing that we're trying to do too is that we're actually calling ourselves an Ai instrument at time or 1 of the products that we're offering is an Ai instrument where it's just another instrument in your arsenal that you can use to create music and so this sort of helps us differentiate from this really vast ai music kind of umbrella that that we have. Where we don't want people coming in expecting that they can remove stems or that they can identify music because we don't offer that where we're sort of but what we do offer is like I said yeah that instrument then kind of keep you moving and then in terms of kind of other software in the space. A really important part for us is that it's not another thing that you have to learn. But it plugs into whatever you use to create. So for anybody creating recording music. For example, you've got this software called digital audio workstations and it's essentially recording software where you have all your nice tracks. Sometimes you record piano tracks then you record.
30:50.20
Jeff
Your voice and drums and then until your song's done and you and you release it so what we do is we we have a plugin that will release soon that just fits in right there you don't even have to open up the browser you just drag and drop the track that you've been working on and it'll continue to write the music there and then you stick it back in your track. So there's very little learning curve. Because these types of plugins are so commonplace. Um, you have tons of other kind of plugins. Not even that are competitors but are just other plugins that exist. So. It's a very natural kind of move to just sort of drag and drop stuff in. So. Because 1 thing that we get when we talk to other musicians is they're just like I'm not going to use something if I have to like think about it and work hard at it unless it's something I really really want usually they're just they've already figured out what they want to do at least for the pros and they just want stuff that's going to either. Help their workflow and increase productivity but they're not going to do something that slows them down. Um, so so there's that route of it and then there's also the more old school route and the one that I did a lot in university with classical music which is notation software so writing in all the notes. It's kind of like a word. Microsoft word. But for music where you're writing out all your scores and all of that right now we just have a where you would have to use the browser still to work with that where you take your music you drop it into Staccato and bring it back when it it generates new stuff for you but eventually we'll have plugins for that as well where it just sort of.
32:14.70
Jeff
Almost like autocorrects but much more predictive music for it where it can pop up and sort of show you where it's going.
32:22.97
mike_flywheel
That's cool. So today it it exists primarily within the browser for some of the elements. It's going to start to plug in right to your audio um creation software like what are these softwares called just for for people listening if they use it. They're like yeah yeah, that's what I use or that's what I was thinking of using.
32:24.50
Jeff
Yeah.
32:36.98
Jeff
Um, yeah, the most popular ones are logic garage Band The the light version of logic Ableton Pro tools qbase are sort of the the big ones.
32:47.17
mike_flywheel
Got it.
32:48.99
Jeff
Um, and then on the notation software side Sibelius and finale muse score those are the the big 3 although I am seeing a move with younger generations of browser related applications. So we we don't see it sort of dying off in the browser. We're just trying to. Make sure everybody can create where they want to create.
33:08.63
mike_flywheel
That's cool so plugins um, generally for a bunch of these platforms are they currently available will be will be soon. This is like ah a current roadmap type item.
33:17.60
Jeff
Yeah, we be ready very soon. It exists and I play with it and I like it I think it's fun. Um, but it's just a matter of getting some infrastructure around it to deliver it to people and make sure all the bugs aren't there like I said like I. I develop it for myself I can always use it and it's always about make sure other people can just as easily use it exactly.
33:38.17
mike_flywheel
Yeah, you want to minimize frustration even for your your early users. Um, you're kind of talking about 2 things there and maybe maybe you can unlock that for me, it sounds like there's this element to help writers block and the writing of words in music. And then there's this element you were talking about the actual music I guess the notes or the instruments or I'm not probably using the right terms but there's 2 elements to the music and staccato actually has 2 parts. They're two separate apps all in 1 how does that work.
34:08.50
Jeff
Yeah, so I should clarify so we we have the lyrics we have 2 apps essentially on our platform. So we have the lyrics app and then the Ai instrument music writing app.
34:12.42
mike_flywheel
Okay.
34:21.76
Jeff
So the the lyrics 1 works similarly to stuff that you might see if you played around chat gbt. But what we bring is you don't have to prompt. It. You don't have to write all these like write me this and that and figure out the best way to sort of get the ai working. We'll take care of that for the songwriter. And all you have to do is basically pretend like you are writing your song as normal but you have more options on the side in case of you get stuck. You're halfway through a verse and you get stuck. So then you can use these little buttons on the side and says just generate 2 more lines for me based on what I've already written. So it'll still rhyme according to your rhyme scheme. It'll still pick up all your themes and stuff like that or if you just can't come up with any chorus any bridge again, you can either pop that straight into it and it'll generate a chorus potential chorus for you or you put in keywords based on what you're thinking because. What I find and I think a bunch of other songwriters find it too is sometimes you have enough of an id in your head but you just can't communicate what that actually looks like so if you just write in a few keywords sunny day hanging out with friends stuff like that. It'll write an option for you and you can either take that whole chorus if you like it or you can just scrape out the stuff. You don't like but at least it's hopefully enough that it gave you that spark to go oh. That's what the chorus is that's in my head. Okay, that's almost it and then I'll I'll change it from there so we just try to give really simple application of this.
35:51.84
Jeff
Kind of cool cutting edge Ai technology where you are just working really fast and you don't have to sort of get out of that creative state of flow that we're trying to always keep you in and then the the music app. Oh sorry yeah.
36:01.50
mike_flywheel
That That's really oh yeah, no go go ahead I was just like mentally seeing it So it's like it's like an assistive hey while you're writing along the side is producing future options of things you could kind of color in against what you're doing for the the lyric app.
36:15.17
Jeff
Yeah, exactly and and to differing levels so either a whole chorus or a couple lines or even just a couple of rhymes that might help you out so just depending on how much help you need. It's it's it's there for you.
36:29.60
mike_flywheel
Got it.
36:32.41
Jeff
And then yeah and then the music side the Ai instrument works kind of in a similar way. Um, we use what's called Midi and it's a protocol that if you've ever seen a keyboard where they push the buttons at the top and suddenly it's a piano suddenly it's an Oregon suddenly it's strings. The underlying information that's being passed to the sounds are is this midi protocol. So it's saying the the person hit this note at this ah velocity like how hard they hit it for and held it for this long and then the sounds go yeah that's great like it didn't say what. Sound they want but whatever buttons they push I'll make that sound and so you can plug and play all these different sounds. But the information is the part that we're interested in so we basically help them at that information level and then they get the ability for all the hundreds of dollars that they've spent on the nicest. Oregon sound or piano sound. They can plug all of that in and we're just going to help them pass along that information or help create the new musical information that can then ultimately sound the way they want it and that's the way that we stay out of purely audio because that's the one that makes me nervous where. I think there's going be a lot of lawsuits because you can take all the audio of the world and splice it into little half seconds and regurgitate it and call it a new song but there's still I p in that half a second of somebody that played the guitar for half a second in Abbey Road or something or drums and for half a second and I worry about.
38:06.30
Jeff
Messing with anything around a creator's Ip Even if they just recorded it for that tiny little bit of time so we were just right at that creator level. We don't mess with what's been published and exists as true audio.
38:17.97
mike_flywheel
Guys You're not taking from an existing source. You're kind of just playing with the instruments within the panel if you will alongside the creator.
38:24.22
Jeff
Yeah, and and the information on how those instruments will eventually sound and so of course if you use our thing to mimic a beetle's Melody. That's kind of on you that you shouldn't do that in the same way that I can't just write a beeles' Melody by Heart. And publish it. So It's sort of we're doing everything that we can to sort of combat copyright and we even have a sort of a copyright checker kind of in that turn it in style thing where it will say like look don't.
38:52.39
mike_flywheel
Cool.
38:58.11
Jeff
You might want to revisit this one section. It's looking a little too close to something that already exists or something that somebody on our site already wrote so you might want to rewrite that and so we're doing ever. We can everything we can to sort of keep the creator's mind at ease that copyright is taken seriously.
39:13.86
mike_flywheel
That's amazing. So like like really responsible Ai capabilities being built right in from day zero and in terms of how you're thinking about this.
39:20.31
Jeff
Exactly because Beyond what the pain point of the creator themselves is the the scariest thing for them is that the music that they put in is going to appear in somebody across the world's music and you don't want that So what we really try to do is.
39:36.35
mike_flywheel
So.
39:39.10
Jeff
But everything in place that we can to stop copyright from happening so you can just create and just know that it's your knock your stuff's not going to get stolen on the other side of the planet.
39:49.81
mike_flywheel
That's cool now I talk about pricing and how you make money but the the question that I can't help but think is if I'm using Stacato to help me create either melodies or or lyrics.
39:56.20
Jeff
No.
40:01.94
mike_flywheel
Do I own the Ip Once this is done or is there any copyright risk for me as an artist utilizing this tool to create things or it's all mine as long as I'm kind of paying Staccato in whichever way that is.
40:12.72
Jeff
Yeah, it's yours like you. It's at the it's at the creator level so you haven't even published it yet. So you you figured out you figure out what you want to do? We're going to help you get to your song. And then ah you publish your song and you put it on Spotify and it's it's yours. What? what you created in the end of yours because it might also change from what we helped you generate? Um, so we we really want to make sure that it's just helping creators. And yeah, we try to stay away from. Anything that might steal stuff from them. So so yeah, it's yours.
40:47.89
mike_flywheel
I Love it. So How how much does this cost. It sounds expensive but I'm sure you're doing it for for an amazing price but like it sounds like a super powerful tool if I'm a creator to output more content faster without the block or as an assistant. Um, what are the ways that you price this out and how do you.? How do you in Staccato make money.
41:06.41
Jeff
Yes, so it's all subscription models. We do have um, sort of trial period 2 where you can test out some of the tools before you buy it and then it's yeah, it's way cheaper than if you're going to hire a human co-writer to work with you and and pay them hourly. Um, it can be cheaper than buying all the music books that you could possibly find and sort of figuring out for yourself. How certain bands write their music so we have if you just want to do the the lyrics side. It's $8 american a month and then the music side is. Eleven fifty and then you can combine it for $15 to get to both apps and all american pricing so convert for any canadians um, and yeah, but like I said it's it's it's vastly cheaper than hiring people and sort of buying content or taking online.
41:46.90
mike_flywheel
Yeah.
41:59.55
Jeff
Courses to sort of get better at it and it's a lot faster because it'll it'll take the do the work for you and figuring out what is the next best thing that could come in this song and that's kind of the beauty of the deep learning technology is that it's allowing us to get to this cheaper and faster place for music creativity.
42:14.40
mike_flywheel
Yeah I love it now again I want to just sort of get ahead because I know someone's thinking this like oh this Jeff guy in this stacato I'm just going to use chat gp I don't really know why I would pay for that when I use chad gp can you talk about like why this is different.
42:26.65
Jeff
First where.
42:31.96
mike_flywheel
And what's unique about it versus someone just trying to do like you know, use chat Gpt directly.
42:38.60
Jeff
Yeah, yeah, the the biggest thing is that we come from that musician standpoint first. So when I was back in grad school and I was just writing papers and not necessarily developing any of these these applications. The 1 thing that I was noticing a lot is that a lot of this. Research comes from people that are in computer science first and maybe are serious even like super serious hobbyists in terms of music but they still built their their life around sort of the tech side first and the music side Second. So what I was and not to knock any anybody like that still cool that they did that. Um. But the thing is that what we were finding is that they were forgetting common music theory things to musicians so we would look at that and go ah it's just basic music theory but they were forgetting it because they didn't go that same route that we as musicians took and they were just sort of quickly trying to figure out the the music theory that. Made sense to whatever application they were using it in so we try to bring that same approach to to what we do in terms of I'm a musician I started there so I can bring as much help to that as I possibly can and then I have a lot of friends that I've played music with over the all the years and then I can. Um, draw upon them for what they would want to see in something like this. So. It's always about what helps the musician first and what's going to be the best for their workflow. So yeah, for sure for a chat gbt example. Um one of the best things that we offer is purely just.
44:06.91
Jeff
Do you want to write in questions of Jot Gbt and do you know how to write in the questions that will get the best sort of response out of chat gbt and then also we just train our models based on more music specific things rather than how they train their models on where it's just everything in the world. Of text so we can kind of curate it towards musicians as well and then the music version is exactly the same where there's just in our opinion nobody that compares to kind of the the level that we've trained it and then just how we've trained it in the sense of we're training it so musicians can use it not just so. Ah, computer can maybe write music. So I think that's the most important and unique aspect that we bring to it. It's just musician first.
44:52.49
mike_flywheel
Yeah, they're different tools. It's it's a tool that could get you there but we're talking about the right tool for the right job or the right instrument for the right note if you will um I love it. So Jeff what sort of on this journey has been maybe the the hardest part.
44:56.90
Jeff
Um, exact exactly.
45:09.92
mike_flywheel
Ah, and is there any advice you can share as learnings from something. Maybe that was a challenge and getting to where you are today.
45:15.98
Jeff
The the hardest things are like I said um for me in terms of the server and just making it So other people could use it that was a big hard part or getting caught up on the business side of things. So The biggest advice that I would have is just no one to delineate tasks and whether that's. Um, find somebody that you trust that might be ah, a great cofounder and I got super lucky in that regard. Um, and I learn a ton and anything that I don't know or just never going to get to that same level that he has we would delineate those tasks to him and because it's for the purpose of making the company. Better. So know when know where your weaknesses are and see if there's somebody that can come on Board. It's hard when you're a startup because you don't necessarily have the money just to hire the best talent around but pick brains as much as you can for anything that you don't know buy lunches like there's always some way to get. Some of that information that you don't have um from people around you and ah just yeah I think knowing where the weaknesses are and then finding the right person that can help you as much as possible and if you do have the money to hire the people then get. Somebody. That's just so much more talented at you at that role than than you could ever be and then it's only going to serve the purpose of making the company Better. So That's that would be the biggest advice and from from what I've found and starting from purely just tech and music and then trying to start a company from there is there's a lot of people that I need to help me.
46:47.60
mike_flywheel
Very very helpful advice around kind of closing off your blind spots with amazing talent and even if you can't hire it go out there and ask for help Sometimes it might be free sometimes it may just cost you a lunch but it'll help you get across the finish line.
46:47.72
Jeff
Otherwise it's not going to go anywhere.
46:54.70
Jeff
You? ah.
47:00.70
Jeff
Um, I thought.
47:02.59
mike_flywheel
Um, what's sort of been your most memorable experience or or part of the journey so far.
47:08.20
Jeff
Memorable moments I mean for me, it's all these little milestones I mean just taking something that was created just like in my room and for school and then the the first customer that you have come in are they even just the first person that signed up. Was super memorable and then that first person that actually paid you money to use something that you created initially for yourself that was that was really memorable. So there's just all these little milestones and it doesn't have to be like your company got super rich or anything like that. It's just these little things along the way that. Ah, been incredibly memorable.
47:46.68
mike_flywheel
It's just funny. It's a very common um common thread or theme is that first person that wasn't a friend or a family member that paid discovered you and paid as little as it might have been like that little bit of revenue.
47:53.17
Jeff
Yeah, yeah.
48:02.90
mike_flywheel
Um, everyone always kind of recounts it as spending you know 5 ten one hundred x that on like beer in celebrations because of how meaningful it was to the work that they were putting in so it's cool that you you 2 have reflected on that I love it. Um, the.
48:08.68
Jeff
Yeah, yeah, yeah, a screenshot sent to everybody just through text like look somebody signed off.
48:21.64
mike_flywheel
Months ahead. What sort of what help are you potentially looking for what are some of the big Milestones ahead and if people want to find out more where where should they go or what should they do.
48:31.22
Jeff
Yeah, if people want to find out more. You can just Google Staccato ai or our website is staccato.ai So that's SdaCCAto and you can just. Sort of check it Out. You can sign right up and it won't cost you any money and you don't have to put it in your visa card or anything and then just start playing with it and that's kind of the best way to sort of get in an impression of how it all works. This is just try it out yourself. What was your other question I.
49:00.40
mike_flywheel
Just like in next six months any any major interesting moments for you or things that you know Staccato is looking for for help or support in the next. Maybe you' raising. Maybe it's just about having more people using the platform so you can learn and they can learn and you can grow.
49:18.22
Jeff
Yeah, where our our biggest thing is is basically just scaling everything up and like I said getting more talented people on the team. So yeah, we're in the middle of of raising and um, hopefully that will sort of conclude over the next few months and then in late. This year or super early of next year will we'll have hired the new people and we're just going to churn out all of these features that are right now just sort of in the project backlog and I'm really excited to bring out there and it just is a matter of sort of having more hands on deck to make it happen.
49:51.72
mike_flywheel
That's cool and what what? what raises is is a preseed seed round where where you at prison. amazing I love it well amazing work Jeff you you and the team at staccato are doing some really cool stuff I love your approach.
49:53.52
Jeff
Pretty yeah yeah Presea So we bootstrapped everything up to this point and it's now it's time to expand.
50:08.59
mike_flywheel
And thinking around how you want to empower creators rather than replace creators and a lot of the work you're doing around responsible Ai and and building it in right from day zero. Um any any closing thoughts or closing words on on your side before we kind of wrap up.
50:23.96
Jeff
No I think this was great I hope that ah it was interesting to to learn about what's happening in the kind of the ai music space and and I hope yeah I hope people will check it out and we're always looking for feedback or ideas on on other features. So I'm always happy to receive feedback.
50:37.40
mike_flywheel
Amazing. Well thanks for joining us Jeff thank you everybody who tuned in to this week's pitch please podcast and make sure to catch us on the next episode have a great day. Jeff.
50:45.55
Jeff
Thanks a lot Mike this is great.
![Unleashing Creativity: How AI is Transforming the Music Landscape with Staccato](https://img.transistor.fm/kjF6IPmvH07EcdWBqxkb-OZaWJzLdKT-xgTUbrEGBjo/rs:fill:800:800:1/q:60/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9lcGlz/b2RlLzE1MDY5NDMv/MTY5NDkxNjcwMS1h/cnR3b3JrLmpwZw.webp)