Riding an AI Rocketship with Mikey Shulman, CEO of Suno
This week, we're publishing a conversation that we had with Mikey Schulman of Suno. So on the website, we have this idea that 10 person company will do a hundred million in revenue as kind of a generational four minute mile for today's entrepreneurs. This was a funny conversation because it was originally presented as one of these kind of four minute mile examples in his journey with Suno towards that end. And it took some really interesting and unexpected turns. We did talk about that.
Speaker 1:We cover it. Their experience is incredibly unique in terms of going totally vertical with growth at the launch of Suno. We cover the impact of AI on creativity broadly and music specifically, and we do get into their journey, the fundraising that's now totaled upwards of a hundred and 30,000,000 from some of the top investors in the category. I think there's some really interesting lessons to tease out of this and stories that Mikey shares. What I probably took away from it most wasn't anything to do with the fundraising, wasn't anything to do with that four mile stuff.
Speaker 1:It was really around the kind of dance that is happening now between humans and AI. And I think it's easy to be incredibly dismissive of AI's role in creativity. The term slop gets used consistently when it comes to AI and creative endeavors. And yet what I found in Mikey was, like, a really soulful person exploring a category of creativity that he has a deep connection to and how the technology they're building can not only create new connections, but also draw out areas of creativity from a human. They might not be able to find their way there on their own.
Speaker 1:I think it's something that I've been sitting with a lot since we recorded this, and I hope you not only connect to those ideas, but that there's others in here, that are relevant to you and what it is you're exploring in your own journey. And so we hope you enjoy this week's conversation with Mikey Schulman from Sufa.
Speaker 2:Never comes out well and then the recording. These demos never go well. Here, it's you know, this is the
Speaker 1:Oh my gosh. When did you get this? I don't know. A couple weeks ago. It's
Speaker 2:Do you want something to drink, by the way?
Speaker 1:What do you got? Do think I'm Diet Coke?
Speaker 2:Both of you. Yeah. We have Diet Coke.
Speaker 1:Come on. Zero or diet? Zero. Right. But I rarely ask for zero because there's not enough people.
Speaker 1:Stop. Look at this. One of the people who flagged me for you was TJ over at Matrix.
Speaker 2:Oh, you know TJ?
Speaker 1:Yeah. We've known each other for a long time.
Speaker 2:He's cool.
Speaker 1:He's really cool. We had this thing on our website, and he would always like tease me about it. It says something along the lines of like, a hundred million dollar company with 10 employees or less is this generation's four minute mile.
Speaker 2:Okay. Okay.
Speaker 1:And so he would like go back and forth and lean back and forth and lean. I think he was still getting settled in over Matrix and then he was like, I actually think Sue knows that company.
Speaker 2:We're more than 10 of one.
Speaker 1:You're more than 10 now, but like
Speaker 2:We were very lean at the beginning, but now we're less lean, but that's also okay.
Speaker 1:I mean, you raised a hundred and 25,000,000.
Speaker 2:We did raise we did raise a big pile of money, yeah.
Speaker 1:But that wasn't your first raise, right?
Speaker 2:No, no, no. That's like the total. Oh, one
Speaker 1:twenty five total.
Speaker 2:Yeah, it's like a little north of that, the press kind of got it a little bit wrong and and we just didn't care to correct them.
Speaker 1:What's the funding arc here? You you did
Speaker 2:Well, it's a seed, an a and a b and then it sums up to a hundred 25, but we, on purpose, didn't announce it until we wanted to, like, hire very quickly. And so then you announce it all at once and it looks very impressive.
Speaker 1:So what was the team like at each stage? So seed, was usually
Speaker 2:I don't even remember.
Speaker 1:Are you saying, tell me, it wasn't that long ago. No, I know, but it's all a blur.
Speaker 2:We were 12 on January first of twenty twenty four, and we're 56 There's
Speaker 1:a lot more here. Yeah. Yeah. How do all this kind of come together?
Speaker 2:Speed is a really funny story. We pitched a different company. We didn't think it would be this easy to do good generative audio in general, so we were pitching more of an audio understanding tool, and then we were thankfully wrong about that.
Speaker 1:What did that mean, audio understanding tool?
Speaker 2:We kind of realized that audio was very far behind images and text in the world of machine learning and we thought that it would be very hard to do good generative stuff at the scale that we were Were you
Speaker 1:trying to use something and you're like wow this is actually pretty terrible or
Speaker 2:is it the world's gonna So at our last job we were doing all text until we stumbled on, so we're at a company called Kensho, machine learning for financial services. We're acquired by S and P Global and we're doing all text except for this one project which is like the least sexy audio project in the world that's automating earnings call transcription. So if like if you've ever read the transcript of an earnings call, very good chance S and P did it. Long story short, successfully brought a lot of speed and scale to that process with machine learning, but we just realized like
Speaker 1:And you were transcribing audio to text?
Speaker 2:Exactly. Okay. Exactly.
Speaker 1:Got
Speaker 2:it. We just realized that it's so Audio is in the dark ages in machine learning, like there's no tooling around it, there's no capabilities, the technology is very backward and images and text are getting modernized and revolutionized and nobody's doing this in audio and we thought we could do it better and so we set out to do that.
Speaker 1:Before you started, did you theorize on why that was the case?
Speaker 2:Like why wasn't anybody trying to do Yes. We I have I have several theories. I don't know if you can prove any of them. It's annoying to do audio. Audio is very big.
Speaker 2:The data is big. It's really messy. So it's got to be a labor of love to to to and you know, thankfully we like audio. And so the other thing is like the dollars aren't there. You know, and if you want to be like really mercenary about it, a ton of progress in machine learning has happened in the big industrial labs like Facebook, Google, now OpenAI.
Speaker 2:Traditionally, like you know, for for a decade it was all driven by what they can sell ads based on. You get sold ads you know based on what you read and based on what you look at and until only very very recently does anyone sell ads based on what you listen to and so there isn't a good reason to do research in
Speaker 1:in all so interesting. I don't know, but that
Speaker 2:makes Yeah. We got lucky that it turned out to be easier we could do it at a smaller scale, than we thought we would need. A lot of people said go build a speech company because speech is like a huge part of audio. What we would do is we were like building these speech things during the day and then we would play with these things at night making music and at some point it was just like what are we doing? Like why are we
Speaker 1:allowed Is that what you pitch? You pitch the speech things
Speaker 2:in your seed No no, we pitched the general understanding tool and then basically right after the seed round we realized it was far easier to do generative stuff at a smaller scale than we anticipated. And so this like I I remember exactly when it was because it was right when Silicon Valley Bank collapsed. That's when
Speaker 1:we were out pitching up
Speaker 2:the That's
Speaker 1:a tough thing to forget.
Speaker 2:And interesting stuff. You know, it's actually fine if you're a company with no money in the bank. We pitched basically and we said like, we're gonna be able to do x y and z right now and then we're gonna you know build the business, get revenues, raise more money, get more compute, make these models much bigger and that's when good generative capabilities were will come out. We basically just got skipped to the end of it Wow. Because it turned out to be easier.
Speaker 2:Was like just a lucky accident. I think our timing our timing was really lucky here. You know, like six months prior maybe there wasn't enough open source software that we could borrow from the text world in order to make this easy.
Speaker 1:Oh, interesting. Say more about that. What did you borrow from the text world?
Speaker 2:One thing I point to that is hugely inspirational is like repository called, Nano GPT from Andre Karpathy. It's just like, this stuff doesn't have to be scary or magical, it's just like, it's just code. Slowly the tool chain got built out for people making these large text models and we didn't have to build out quite as much, know, PyTorch having, something called FSDP, you know, which lets you train bigger models more efficiently, so stuff like that. So six months prior we wouldn't have been able to do it, we would have had to do a lot more by ourselves. Basically six months after people start to enter the space too, so our timing was just really lucky that we kind of threaded the needle there.
Speaker 1:And was your seed round the starting gun or had you been working on this for a while, did you have something
Speaker 2:to walk into We were working on it for a little while. I feel very lucky, who we have around the table. Yeah. People I really really like. I'm a normal people person, you know.
Speaker 2:Know, life is too short to work with people you don't like.
Speaker 1:If you go back to that first pitch, how big a vision were you pitching? Like how big did you think this thing could be?
Speaker 2:Actually quite big.
Speaker 1:You did, even back then.
Speaker 2:Yes. Okay. Because if you just think about how much audio is out there and created and just underutilized and I thought it would take a really long time and still will take a really long time for everybody to utilize all the audio they have, but like, you know, your ears are like always on, you know, there's always stuff happening and it doesn't get utilized. I am actually quite bullish on audio in general, like outside of music. There's just tons of audio out there that you know, eventually, there are some futures, you can think of them as dystopian, I'm actually not sure how I think but just like everything's always getting recorded all the time like people are walking around with recording devices and in this future there's even more audio than there is today.
Speaker 2:So I think audio is actually really big. We now focus on a narrow but really fun slice of audio called music that I think is truly wonderful, like it makes me very happy but, there's a lot more to audio than just music. Okay. There's a lot of people doing good stuff in audio.
Speaker 1:Yeah, how has that changed then post, like you said, six months earlier, six months later you timed it perfectly, what does audio look like today versus when you started?
Speaker 2:Now there's a lot of stuff in the kind of productivity and sense making. You can think things that are adjacent to reasoning, so all of the big LLMs, OpenAI with its voice and Anthropoc, like all of these things now have voice as an important modality because, or audio as an important modality because that is kind of how humans communicate. Thing you can think is, we kind of evolved to communicate through audio and writing stuff down was a hack that turned out to be really good, but I think it's very hard to just completely get rid of it. And I think a lot of companies that went all remote and now don't do all remote, I think kind of feel this that there's a lot of stuff in passing that gets done, human to human that's like that. I think in general, there's tons of stuff focused on productivity gains which is really really important for growing the economy and one thing we say here is like fun is underrated and it's like the whole world is focused on productivity with AI for very good reason, but like let's not forget that fun is important and being fulfilled and enjoying yourself is really important.
Speaker 2:You know, I like to think that you know all of these productivity gains are basically making more time for you to have fun for example with music.
Speaker 1:That's a great way to think about it. But most people don't. That's right. You know, in fact, you talked about dystopian futures. It feels like most people are really swirling around dystopian futures versus inevitable futures that have massive positive gains?
Speaker 2:Yeah, think so. You know, maybe this is like a bias of the human mind that we that we glom onto the negatives, more than we do the positives. Know, there's like a realization that most sci fi futuristic movies are dystopian not utopian. I'm a firm believer in human agency and we can build a good future and not a bad future for music, a, but actually just in general, and so it need not be dystopian.
Speaker 1:How has your views on humans changed now that you're working more with artificial intelligence?
Speaker 2:I'm not sure, I'm not sure.
Speaker 1:If it has at all.
Speaker 2:I'm not sure it has changed. I the thing that makes my views change the most about humans is being a parent. Oh, That is like by far makes you like just completely change what you think of.
Speaker 1:You've got how many kids? Two. Okay.
Speaker 2:Yeah. Come from a very big family.
Speaker 1:Do really? How many?
Speaker 2:I have three siblings but I have 50 something first cousins. I have a very big family. Really? I love it. Wouldn't have been Who was the
Speaker 1:overachiever in that family? A lot of overachievers.
Speaker 2:Everybody's an overachiever in that family.
Speaker 1:That's incredible. Wow. Was music always a part of the family?
Speaker 2:Yeah. So I started playing when I was four, but I have two older
Speaker 1:brothers. Just started with just piano? Yeah. Just piano. I don't play anything.
Speaker 1:So this is exciting.
Speaker 2:I have two older brothers. So, you know, there was definitely always like an example of someone better than you just always in the house. At some point I picked up a bass maybe 12 or 13 and played that much more aggressively and much more you know I would actually practice and so I got better quickly. Played in a lot of bands in high school and college
Speaker 1:What kind of music?
Speaker 2:Enjoyed it.
Speaker 1:Kind of everything. Oh, really? Okay. You didn't have one genre.
Speaker 2:No. A lot of like rock and jam band adjacent stuff. Okay. But tried to dabble in everything. Yeah.
Speaker 2:Like music is cool. Music is special. I think there are formative periods in your life and the music you happen to be listening to then ends up gaining importance and maybe it's happy back Like a lot of fish. There you go. Okay.
Speaker 1:Were you growing up around here?
Speaker 2:I grew up in New York. Oh, did? Okay. Yeah. So, so I've I've I've seen them many, many times.
Speaker 1:Yeah. So what's been surprising you so far growing Suno?
Speaker 2:I think how much fun it is. Like, yeah, sure. Building company is hard, but it's also really fun. And I think that I feel I feel very lucky that I love my job as much as I do. I don't know if everybody has that.
Speaker 1:Well, mean not everybody's strapped to a rocket ship.
Speaker 2:That's true. That's true. You know, but you know, in some sense like some of the least fun things also come from, being strapped to a rocket ship.
Speaker 1:How so?
Speaker 2:I love it. Okay. You know, maybe I don't know. Don't have a counterfactual. I don't have a counterfactual.
Speaker 1:Well, you've done Stardust. You were at Kensho?
Speaker 2:Yeah. That? Kensho was a rocket ship too.
Speaker 1:It was?
Speaker 2:Yeah. Oh, fascinating. So, and a very entrepreneurial bunch. A lot of people have left, have gone on from Kensho to start companies.
Speaker 1:We had this four minute mile thing out there. You're seeing so many companies getting to like really massive scale in your world really, really quickly. You've lived some of that. What it looked like going from like audio researching, there's something here we're gonna figure it out and your seed round to like, you know, Mike, not just Mike, but I know that was like a really competitive series A that you raised. Like
Speaker 2:Productivity is good. Growth is good. I think most people don't really appreciate that. And I feel very fortunate the arc that we were on.
Speaker 1:How did that arc emerge though? Like you didn't just You
Speaker 2:don't even realize it in the Stop it. No you don't, you don't. We're not in Silicon Valley, we are not as steeped in AI, AI pilled as many other companies. Just as an example, whenever something that feels kind of apocalyptic happens in AI, whether it's the weekend Sam Altman got fired and rehired or the weekend that Deepsea comes out, we have a much more regularized, to use a machine learning term, we have a much muted response to these things and it's just, it has pros and cons.
Speaker 1:What do you mean when you say muted? We'll have like
Speaker 2:a normal work week, we won't all be, you know, everyone will
Speaker 1:Nobody's you know, just like
Speaker 2:Nobody's doom, not like, you don't have 80% of the company doom scrolling, right, And that's really good for focus. You know, it has it has its downsides for sure. Which would be what
Speaker 1:though?
Speaker 2:The downsides? Yeah. Oh, you wanna be current and up to date. I think I think the thing that people get wrong is thinking that I need to know everything or it depends on who, but I certainly don't need to know everything the instant it happens and if I find out about something one or two days later and it gets kind of summarized or digested for me by the world, that's actually probably a good thing. I need not panic, like certainly I, the CEO of Suno need not panic the weekend that Sam Altman know departs and comes back to OpenAI or I'll give you an example, I remember I was talking to a friend on the phone one evening and he says to me, hey, I gotta go, there's, GPT-four launched last week and there's an emergency hackathon, this evening so I gotta go, I don't wanna be late.
Speaker 2:And I remember just thinking to myself like, what the emergency, right?
Speaker 1:Now you sound like a dad of two.
Speaker 2:Maybe, maybe, right? Like, I wasn't at the time,
Speaker 1:but Yeah.
Speaker 2:So I think that that that focus and it, basically smoothing out the bumps of the emotional roller coaster because it is a roller coaster doing this.
Speaker 1:I bet so, I mean you're lit, I mean it's gotta be a constant barrage of things that arguably are probably just adjacent but certainly like this category is like a fire hose.
Speaker 2:Yeah, totally. But focusing on the things that are gonna matter and limiting distractions is really, really good.
Speaker 1:How do you tune that filter for yourself at least?
Speaker 2:Staying off social media really helps.
Speaker 1:Oh, interesting. Okay, so that's not part of your diet? No. Okay. Is that a new thing?
Speaker 2:No, it's always been. In fact, I got Twitter this year or last year because people told me I should and I don't have Facebook and I don't have Instagram and and and I don't know why I'm the right age to have
Speaker 1:All of that Yeah. All
Speaker 2:of these things, I didn't. I I feel lucky that I never really got into social media. Having the right people around me who are the right positive influences both at work and at home is like hugely, underrated and positive effect. An unnamed very success successful venture capitalist who is not on our cap table told me once he's like, really underrated is how stable is somebody is an entrepreneur's home life. Yeah.
Speaker 2:It really struck me. You know, life is certainly busy at home. I have small children, but it is wonderful and stable and I think I feel so blessed that I have that.
Speaker 1:Yeah, I feel the same way. Mean, just had our twenty ninth wedding anniversary. Congratulations. Thank you. It's a superpower.
Speaker 1:I mean, really genuinely is and I don't know how you felt that here.
Speaker 2:Mean, I feel it every day, right? It's like it's a calming and normalizing force in your life, right? Work is certainly fun, but a hectic and having, yeah, just enjoying your home life is very important.
Speaker 1:How did the office culture here start to take shape? Because obviously you've got your family, but you also have what appears to be a pretty intense work culture too.
Speaker 2:I don't know about intense. It's in person, so, you know, that's not for everybody.
Speaker 1:Yeah. So, no remote offices.
Speaker 2:We have some people who are remote, but the vast majority of people are in the office. We actually have three offices. We have a office in New York and we have an office in LA.
Speaker 1:Oh, amazing. How long have those been around?
Speaker 2:New York's been maybe five months and LA is actually like negative one month. It's going to open very soon. During the pandemic when we were at our last company, I would actually go in after like things calmed down, I'd go in every day and there'd be like three people there. But I think people just don't realize part of the reason that people were doing in person work for a while is like there was no such thing as Zoom, but there are other reasons that just got wildly overlooked. And I think there's the same thing of like everybody kind of just going all in on things without thinking about them whether it's like deep seek or whatever other flavor of the week it is, it's like I feel the same way about remote work which is that it has its benefits but it definitely has its downsides and we would be unwise to overlook those downsides.
Speaker 2:This is a team sport, this is a contact sport Mhmm. And I wanna be around people. And this is a personal thing. I'm like, I really get energy being around.
Speaker 1:Well, I think this is the this is the thing that I said early on in COVID where people were like, oh, remote's remote's the way to go. And my take is always like, whoever's leading the company, whatever they're inclined towards, like that's what's going to work. So if you think, if your CEO hates remote, sorry guys, like remote's just not going to work. Like you're just not tuned to work that way.
Speaker 2:That's right.
Speaker 1:And you've got to set it up in a way and surround yourself with people that are on the same wavelength as you. Part of what you said is kind of blowing my mind, which is like you run a consumer app and you are not on social media. Like that's pretty freaking wild to think. In this day and age, I'm just, I guess I'm just like sitting with that as a thought. I don't even know that there's a question there.
Speaker 2:Like, you know,
Speaker 1:there's so much like hacking and things that like people are trying to figure out on social and how to go viral and get network effects and all that stuff. Here you are with like none of that is context and yet.
Speaker 2:Well, there are people here who understand these things better than I do. We have a chief product officer who formerly ran all the product at Snap. The other thing is, like, you shouldn't need to hack your way toward pleasurable experiences around music. I think humans evolved to resonate very strongly with music and in some sense you just, like, you need to iterate and you need to build it, but there isn't going to be some, you know, I don't think we're going to do, ephemeral songs for people in order to, you know, kind of gain some virality. Like, this isn't, you know, this is behavior creation, which is hard, but it shouldn't shock people that people love music.
Speaker 2:In some sense, maybe it's not so crazy.
Speaker 1:But there's so many been so many cuts at different attempts at music. Why do you think this one worked?
Speaker 2:Some of it is the timing being lucky.
Speaker 1:Lucky, you know, we talked about that a little earlier, just on the tooling. Yeah. Okay.
Speaker 2:Yeah, well like Was there any other
Speaker 1:thing else going on in the zeitgeist that like made it AI was very popular.
Speaker 2:Okay. And so that definitely helps. And you know, many of our initial users, especially like, you know, at the beginning when the music is not good, you need to be, let's say, forgiving of of of, you know, people will still come because it's like, it's a bit of an AI party trick at the beginning.
Speaker 1:Totally, it's like real novelty.
Speaker 2:Yeah. And then if you can get better quick enough they'll stay because it got better quick enough. So we we think about a few things with music. I think, one is the the audio quality, does it sound like crisply recorded? One is the song quality, is it a catchy song that resonates strongly with me that I want to listen to over and over again?
Speaker 2:I'm sure there are songs that you have listened to hundreds of times in your Yeah. And then the last is, we say control, which is like, I have ideas in my head, how quickly can I get them out? And it's like if I ask for a saxophone, did it come out as a saxophone? If I ask for classic rock, did it come out as classic rock? Something like that.
Speaker 1:Of those three, which is the hardest?
Speaker 2:On the because in my
Speaker 1:mind control, like that's such a great metaphor for it, that seems to be the hardest I would imagine so.
Speaker 2:Right? Forever. At a certain point the audio quality will get so good you won't you won't realize. I think so I think there's a lot of room to grow though in in song quality. I think music has stopped really evolving, harmonically.
Speaker 2:I'm quite optimistic that AI is a means to have it start to evolve harmonically again.
Speaker 1:Okay, say more about
Speaker 2:that.
Speaker 1:Quality, harmonically.
Speaker 2:You look popular music in the last decade and it's basically the same song structures, the same chord changes, there's only like three approved topics for songs anyway, there's like love songs, hate songs, and the only where it evolves in an incredibly interesting and quick way is the sounds. Like so any and and and because sounds they hit you, right? And so there's this like these like crazy stats, like if you look at the last twenty years, so twenty years ago, know, maybe a third of Billboard number one songs had a key change in them. Fast forward to 2023, it's zero. There is no reason that we can't go back to things being harmonically really interesting.
Speaker 2:When a key change hits you, when a chord that you weren't expecting hits you, it hits you. Oh, And so I'm quite certain that we will, that these are like kind of natural flowing things and AI is a
Speaker 1:If they know it hits you, why does it go to zero?
Speaker 2:It is just easier to get people to want to listen over and over again if the sounds are really interesting and you can kind of compress all of those interesting sounds at the beginning so people listen to it. And a key change, you need to develop a lot of, a lot of motifs in one key and then switch to another key that takes time and like the attention economy is short, the way streaming is set up, all I want you to do is listen past thirty seconds and then it counts as a stream. Like there's lots of
Speaker 1:pressures that Can you unpack some of that too?
Speaker 2:This is studied a lot but there's basically an element of like for things to be really popular they need the right mix of stuff that's familiar and stuff that is different. And so like as an example, a sequel really gets you that, you know the backstory, right? I guess the musical equivalent might be a cover. People with great voices they always tell them to start doing covers, like start with covers, right? You get that familiarity for free.
Speaker 2:And it's comfortable and it's easy, but it's, where you take it from there is is I think the real question.
Speaker 1:How much do you think the music services and the algorithms then drive that idea of like doing covers or doing the familiar? Do you know what I mean?
Speaker 2:I don't know the answer, but so I think people don't appreciate the extent to which and these things are intertwined, the extent to which both recommendation algorithms and just the general incentives in the system matter a ton. The general incentives in the system, for example, are I'm trying to get streams on Spotify, that's how I'm gonna make money. A stream counts, once somebody listens past thirty seconds, I gotta make sure I get the hook in to get somebody to listen to thirty seconds and then maybe I stop caring as much. Stuff like that ends up mattering a lot and then if you are YouTube and what you are trying to do is get people to stay on your platform as long as possible instead of watch as many videos as possible, like whatever the incentive structure is in the system ends up like kind of in the product and I think that there are no there are no free lunches here, everything is trade offs and so, I think about this a lot, like it's not really in our platform yet, but I think about this a lot.
Speaker 1:Why isn't it in the platform yet?
Speaker 2:We're not really doing like serious recommendation type stuff, know what we're giving people is the experience of creating the music that's in their head and letting them kind of take the reins. But who's to say we won't at some point be recommending music to people in a much more serious way? I think this is also though a very natural progression that we are the next step in, so the last big wave in how music was produced was going from analog production to digital production, you used to need a giant console and you would need to hire tons of session musicians and it would cost a million dollars to make an album production.
Speaker 1:Rip Rubin doesn't know how to use any of it.
Speaker 2:But Rip Rubin got to have people do it for him, right? No, but this, you know, I think I think that's actually a perfect analogy. Rick Rubin, you know, can have people actually do all of the stuff for him and then digital production comes and you can have 14 year olds with their laptops produce hit songs because they will learn the tool, it's all contained on their laptop, they have really good ears, they learn what sounds sounds really good, somebody gave them all of the things that Rick Rubin has which is people who do the things for you or they learned it. And so that greatly reduced the barrier to make great music. You have people, you know, Martin Garrix was 14 years old when he gets discovered and I think about this as a continuation of that where you need even less emphasis on skills and more emphasis on your ears.
Speaker 2:It gets easier and easier to create and what that means is that you rely more and more on your ability to say I like that, I don't like that, how do I tweak that, keep going. It makes a lot more music. Yes, it will make some more bad music, but it will also make much like the highs will get higher. I understand why that's scary for people, but that's really good.
Speaker 1:Why do you what do you mean when you say the highs get higher?
Speaker 2:The best music will will get even better.
Speaker 1:Oh, on what dimensions?
Speaker 2:All the dimensions. It will take time. It will take effort. But if you make it much easier for people to make stuff and experiment with stuff, they will arrive at better music faster. Somebody should check me on what I'm about to say, but my my my
Speaker 1:We cut it out.
Speaker 2:My observation is that it it it comes to younger people first and that is because at that point older people aren't as digitally native and at this point I think most people are most people know how to use an app on their phone or open up a website and I think the big difference here.
Speaker 1:But you have a lot of people mad at you.
Speaker 2:Yes, we do. Is there
Speaker 1:a common theme to what you think they're getting upset, A, and then B, maybe what they're getting wrong about directing that anger towards you.
Speaker 2:There are a few common themes. You know, I think in general, change is scary and technology has not always been kind to the music industry. So I, you know, fully understand I
Speaker 1:mean, I guess Napster wasn't Napster
Speaker 2:was not You know, piracy But it
Speaker 1:seems like the streaming services are beyond kind to them. Right? Like Yes. But it took
Speaker 2:a long time to build back up. That's true. Were expensive and people bought them a lot, you know? And so and Vinyl literally almost died. I totally understand that.
Speaker 2:You know, I also understand people are afraid that when you put the means of production of music into lots more hands that that will take music in in into new places and I think that there are, for good reason, cultural gatekeepers who who who sometimes get worried about that. But I also think that putting the means to make music into more hands is very obviously good. It's like you are letting more people be expressive, you are getting way more people into music and you know you zoom out all the way and you think about this without some of the details and you say a billion people way more into music than they were before, that's gotta be good for music on the whole.
Speaker 1:What's the argument against it?
Speaker 2:I I think then you start to get into little details and you find and you find little places where it's not perfect. I think that the music industry is way smaller than it needs to be and that if you make really fun engaging meaningful experiences for people around music that they don't have right now they will love them and the music economy will grow and more people will be able to make their livelihoods making music.
Speaker 1:What's the most compelling case though against it? That's like, you know, America, apple pie, the whole thing, the way you just said it. Yeah. There's a part of me that's like, well, who would disagree with that and why and clearly there are plenty of people who disagree with it.
Speaker 2:I think so. I think it's everybody who is doing well in the current system is not going to want to see a system change and I totally understand that. There are very common complaints I think that, in some sense are not maybe self consistent, but it's like we are simultaneously making terrible music and ruining music for everybody, and I think that or simultaneously making terrible music and replacing artists. And to me these are inconsistent like they can't both be true like if the music is bad we can't be replacing artists. You know this would be a very damning critique I think but and I am the first one to acknowledge that when you let way more people make music and you make way more music out there, there will be more bad music.
Speaker 1:I just don't understand how people could see that that's like a net negative.
Speaker 2:I think imagine you have spent a lot of time and energy and money learning how to make music. But wouldn't that make your
Speaker 1:music that much more value? Like, you would stand in such sharper contrast?
Speaker 2:It might or it might not. So so I think, you know, so I think it's, look, it's it's hard. It's hard. And and this is this is not new. This is this is the case in writing.
Speaker 2:This is the case in movies. This was the case with hotels and Airbnb and cab drivers and Uber and, and I understand that in each of these cases, there are people who, are gonna be negatively impacted. But I think that the future where just far more people are into music and participate more. I think a lot of a lot of this is a lot of the downside is failure of imagination actually on people's parts of what what could it be, because I actually think what we do with Timberland. So we recently had this contest where he released a song on Suno.
Speaker 1:I saw that. Yeah. And So cool.
Speaker 2:Yeah. And I I don't know if we're public with the numbers yet, but in inordinate
Speaker 1:number of Inordinate number of people.
Speaker 2:Yeah. Okay. I'll I'll say it. A hundred thousand people remixed that song all of us. And when I think about, like, what super fandom can be, this is now co creation with your favorite artist.
Speaker 2:This is amazing. And we didn't charge fans for the right to do that. And Timbaland isn't feeling threatened by it. He is so good at it and so pro AI and so he's like, I think he gets it and he sees it. Yeah.
Speaker 2:It's the future of, you know, of of what can be. Right. He's also so good at, you know, just watching him.
Speaker 1:I mean, but it's an interesting trend because people have flirted with this idea for a while. Right? Like I remember,
Speaker 2:I don't know
Speaker 1:if was Kanye West or somebody else, they released like stems Mhmm. Of songs as they were like leading up to an album and and like really encourage people to go out and use and remix and all that stuff.
Speaker 2:It is an incredibly high form of flattery. I think that people don't realize how special music is when you can play with it and, you know, you release stems is amazing, but you still need to be, like, fairly proficient with the tools in order to really read it.
Speaker 1:Yeah, I mean this stuff wasn't
Speaker 2:around when that happened. If I can take one of, my favorite artist songs and remix it into a different style, adapt it somehow, cover it somehow, sample from it somehow and take my take on it, and show that to people. I think this is like, this is the sociality that is missing from music right now, and it is beneficial for artists and beneficial for fans.
Speaker 1:But people are incredibly threatened by it. Yes. And as an artist, Timberland may be embracing it, but I would imagine there's a bunch of artists who are like, you're stealing my stuff. You're making crap music out of like, my stuff is your inspiration. They're feeling like it was their art, you've now turned it into something that's unrecognizable to them or even offensive to them.
Speaker 2:Well, I'm not saying every artist must let
Speaker 1:their family No, no, That's not, I'm not saying it at all. I'm just trying to like frame up like if I were an artist, what would be my counter to that or how would I think about that? And I would imagine you've got to navigate that too because like Metallica sued Napster out of existence but their stuff's all over Apple Music.
Speaker 2:I would never want to force somebody to let their fans interact with their work. I think there are plenty of artists in all domains who are somewhat agnostic as to how much people like their work. We obviously want to encourage the legal use cases of the product and they far outnumber the illegal uses of the product But I also think that this is one of those cases where the kosher use cases of the product are actually the ones that are long lasting and I feel like impersonate your favorite artist is actually just not. I feel like that's the AI party trick that goes away.
Speaker 1:That's exactly the phrase I was gonna go back to.
Speaker 2:And I think that I think that we all did this with GPT when GPT first came out and you're like having it write a gangsta rap about you know, going to Starbucks or something like that. Then like the third time you do that, you're like, yeah, this is kinda silly, but that doesn't mean these tools don't have immense value to my life because they do. It's just like it wasn't that impersonation thing.
Speaker 1:Oh, that's so interesting though, but it was also the gateway drug. It was the thing that like pulled people in.
Speaker 2:It no longer is though, I think. Maybe maybe now Brock is doing that again, but, I think this is one of those cases where the the long lasting use case is also the one that that everybody wants it to be.
Speaker 1:You guys charged for it right away, which for a consumer app was like, you know, I don't Strange. Strange. So why? And why was that important to you?
Speaker 2:We put this thing out there and we don't know how much people will like it and, we want to make sure that we are on the right track and so we we always had the free tier and if you wanted to use more you have to pay.
Speaker 1:Where does it stop, and how did you decide
Speaker 2:on
Speaker 2:where
Speaker 1:I
Speaker 2:think never changed it, so truth is somewhat arbitrarily. We said, you know, you can you can use it five times a day and then Oh, okay. Which is quite generous. Yeah. And then if wanna do more, have to pay.
Speaker 2:Yeah. Lots of people are are are daily users of the free tier, recurring. And even back then, and I think we got a lot of people to subscribe in the first month, which told us that we were onto something. And I think How
Speaker 1:much were they how much were you charging at
Speaker 2:Same same thing. We haven't changed anything. Nothing. We we added a more expensive tier on top. We added topic, but like the the core structure of the business model has not changed.
Speaker 2:Okay. It's really nice because it tells us like this is the thing that makes people want to part with their hard earned dollars because I'm giving them something valuable. You could make the argument that maybe we would be bigger if we had just never charged and given it all away, but we also may not be because we wouldn't actually have the data on what people are willing to pay for. And that turns out to be really really important because this is behavior creation and we don't know a priori what fun new behaviors are people actually finding valuable and what things are they just doing even though it's not valuable enough. One of the investors in the company, said, gave this piece of advice which has really resonated strongly with me.
Speaker 2:He said, let me write the obituary of the company for you so that you don't have to. And he said, a bunch of you know like brainiacs who think they're smart enough to sit in a room and just intuit what the best consumer experience is going to be and they get their lunch eaten by a bunch of people who just iterate their way there faster. That really resonated with me, especially in something like this where the behavior creation is hard.
Speaker 1:Right? Why do you think that resonated so much with you?
Speaker 2:No, I think it's true. You you know, like you learn a lot by doing.
Speaker 1:The alternative, like the counterfactual to that was what? The alternative universe that you
Speaker 2:Don't iterate quite as quickly. Try to be much more first principles thinking about the types of behaviors that people are gonna wanna do. That might have led us to never charge for the thing and maybe now we would have a little more usage or a little less usage, have no counterfactual but we'd have no revenues now. Revenues are nice and it seems, I used to joke like, oh yes our revenues are really good for an AI business but if everything is qualified for an AI business this is actually probably a bad thing.
Speaker 1:What did you mean when you said for an AI business?
Speaker 2:A lot of AI companies have not massive revenues, you know, without naming names, and it's a lot of hype, and I am glad that we are building a strong business here. Revenues mean, yeah, we raised a lot of money, but revenues mean we burn it very slowly, and that is a real weapon.
Speaker 1:But you went from zero to like pretty meaningful revenue pretty quickly. Mean, there's been speculation. Like how much money, how much revenue you were using?
Speaker 2:Yeah, they were strong. Where? At what point?
Speaker 1:I mean, at peak. The Like you were nine
Speaker 2:Revenues are still growing. I'll decline to give a number. I think the last public number that leaked was 40, but we're ahead of there. That's what Yeah,
Speaker 1:What did iterating do for you? Like, with each turn of the crank, you obviously launched with something. Yeah. And then you launched something again.
Speaker 2:Let me give, try to give the counterfactual here.
Speaker 1:Yeah, love it.
Speaker 2:Which is, because it exists actually, each turn of the crank produces information and then when companies come along and they try to launch clones of your app, they don't have all of the learnings that came from why did you iterate your way there and you can then launch a clone of our app and we've seen a lot of clones of our app come up and it's actually just harder without all of those learnings to know what to do next.
Speaker 1:So you've seen the counterfactuals built.
Speaker 2:Yeah, we are not the counterfactuals,
Speaker 1:we've seen the counterfactual, That's so wild. How quickly do they show up?
Speaker 2:Oh, there's like, you know, there's US ones, there's overseas ones, there's, you know, like at any given point, there's probably 10 of various sizes. Is
Speaker 1:there a theme to what you keep being drawn to iterate on? %.
Speaker 2:I think, I think, you will see us try to make music much more social. I don't mean Facebook social, but I mean music is meant to be
Speaker 1:shared between people Collaborative. Collaborative. Okay.
Speaker 2:Created not just one person, riffed on. I think you'll see us lean very strongly into that. Those are the most pleasurable experiences I think that people have. I have this I have this, memory burned into my brain of like the first the first real show I played on the base, with with two friends, I was in high school, not a good show objectively, like you know like we're still learning and at the end of the night you know we get, you know, we each get like $30 from the promoter or something like that, but I remember just thinking to myself, was handed this money like this is criminal, I had a better time than the audience, it wasn't a good show and I had so much freaking fun with my friends making music and I think this is basically the feeling that we can give to other people. I wish I could just open our our you know like hello at suno.com inbox and show all of the beautiful things that people write into to us about how music is brought to life for them and you know I can tell you like probably half of my usage of the product is with my four year old.
Speaker 1:Wow. Really? Has it always been that way?
Speaker 2:No. But like as he's gotten older.
Speaker 1:Yeah. What do you make?
Speaker 2:Mostly songs about him in fantastical situations. Yeah.
Speaker 1:That's so cool. That's so cool.
Speaker 2:I think music is also really underappreciated in development.
Speaker 1:What do
Speaker 2:mean? With kids. Like, it's a really important part of it. People go crazy, know, it's like ten years ago everyone's like, you need to learn to code and now everyone's like, no, don't need to learn to code. It's like, you should probably just teach kids music.
Speaker 1:How formal is that for you?
Speaker 2:There's probably some revisionist history that it's like the most important thing. It wasn't the most important thing, know, like I focused on school much more than I focused on music, you know, when I was a kid. Certainly discipline is helpful, but there were people who were more disciplined and who were better than I was.
Speaker 1:When things started to work, who were the first few hires you made?
Speaker 2:The seeds, so we were four co founders, we had all worked together at Kentro, one of whom is my best friend. Very cool getting to work with your friends. Very cool. And then early hires were machine learning and software, some from our network, some from not our network. One of the early hires came inbound after we released an open source repo, an open source piece of software.
Speaker 2:Really made me realize like, the the power of actually being a little bit of a known quantity and then people wanna come to the Yeah. Yeah. Yeah. It got a lot of stars on GitHub so people liked It was again this is a timing thing. Was just like, it was the first like very realistic text to speech and we made it open source and people did a lot of funny things.
Speaker 1:What was your intent when you released it? What were you hoping for?
Speaker 2:If I'm being honest I don't think we had totally thought it through. If I'm being honest we hadn't totally thought it through. We think this is really cool, we put the guardrails in place to make it very very hard if not impossible to clone people's voices and then we released it. And it put us on the map a little bit. It's like people start to reach out, potential employees, investors, potential customers, etcetera.
Speaker 1:You're starting to get inbound. You're now getting ready to launch Suno. How does the composition of the team change based on, I would imagine most of your early hires were just more technical people.
Speaker 2:You're a four person company, you're an eight person company, every, you know, these things don't work if somebody thinks that their job is as narrow as the job of a software engineer at a big tech company.
Speaker 1:This is not
Speaker 2:the way these things work like you just, know, like, Alex Wang from Scale had this blog post like you just need to hire people who give a shit. It's just just that. I don't have like a pithy answer here, I've done a lot of interviews in my life and you end up just developing some pattern matching. I've done a lot of interviews in my life and I think, an uncomfortable thing to say is that interviewing is a skill like anything else and you can get better at it by practicing it from both ends. I asked our HR team to actually count how many interviews I had done and I did over 600 face to face for Zoom Wow!
Speaker 2:Yeah! So and I remembered very distinctly early on realizing that like part of my job is hiring people I remember like I'm getting like totally destroyed because between the time it takes to interview them and prepare for the interview and write up good feedback it's like it's becoming a significant part of my job and I said like you better trick yourself into liking this because it's a big part of your job and then I actually love it.
Speaker 1:What do you love about it?
Speaker 2:People are interesting, even bad interviews are really interesting. Just, I like talking to people and so that's, you know, maybe the biggest advantage I have is just like actually enjoying interviewing people. It's not burdensome. Think I
Speaker 1:have You have to talk yourself into that.
Speaker 2:I think I have four interviews tomorrow.
Speaker 1:I remember watching Rick Rubin do an interview with, I can't remember who, maybe it was Pharrell. And they got talking about AI and they got talking about taste. And Rick Rubin was like, AI will never be any kind of factor in music because it can only look backwards and it has no sense of taste. It has no sense of like what is actually good. It just knows the data.
Speaker 1:That seems like an area you're thinking lot about.
Speaker 2:Day and night. I happen to disagree with that. Humans are also only looking backward at all of the music they've heard and they try to come up with new ideas. AI is a tool. I think the the weaker version of it that I agree with is AI will always need human guidance, like a human to pull out the interesting sounds from it.
Speaker 2:And that's probably true, I think that's true.
Speaker 1:You seem to not fully buy that.
Speaker 2:I I like, is there is there a world where where there's some collective way of pulling taste out from society and not a single human, yeah maybe, but the thing that is different about music, if you're like super steep in AI, it's very easy to accidentally apply everything you know from the LLM space onto music and it just doesn't work. And the biggest thing that's different is that in the LLM world you're mostly trying to go after objective facts. You have objective answers. It's a yes or a no or it's multiple choice or whatever, there is an answer and in music there is no answer And if you had played the Beatles to Bach, you know, would probably have said like, and somehow both of those, artists made timeless music. Yes.
Speaker 2:And I think that because it is subjective and also changing, it is very difficult to take the human aspect out of it, but the blanket statement that AI can only look backward and has no taste, I think I think is just wrong because it actually functions very similarly to how humans do. I'm absorbing all of the things that I know, of the music that I heard, all of the culture that I see out there, and I'm finding some way to express it. And I see I see no no inconsistency with like a human with great taste learning to pull that out of the AI and not doing anything to it, like just trying over and over and over again and the right thing comes out and that is like a beautiful piece of music.
Speaker 1:Have you had experiences like that?
Speaker 2:Oh totally, Totally and and, you know, you can you can say that I had a part in that creation or you can say that I didn't have a part in that creation.
Speaker 1:What do
Speaker 2:you say? I almost wanna say I don't care what you think. Yeah, think that's totally fair. I like that piece of music, it's a good music. I felt like I had a hand in making it.
Speaker 1:I was gonna say, but my sense is you're saying like I did have a hand in that. Whether it was No.
Speaker 2:I mean, I think I did, but I think it's also maybe on some level it's irrelevant, like there's there's gonna be questions over copyright. Can I copyright the piece of music that I just made with that AI? I I that is quickly evolving, what what, what the law says about that, but even without it, it's like external to that it's actually an irrelevant question. It's like I like that piece of music, it was good music, I I felt like I enjoyed the process of making that piece of music and maybe like does it count that I made that music or not is irrelevant. Because That's
Speaker 1:so interesting. But isn't that what copyright hinges on right now? Whether or not you
Speaker 2:can copyright a piece of music, yeah, definitely. So the latest guidance is you need some amount of human input to be copyrightable. And I think this is certainly an improvement over the ambiguity that used to be before it. But there's lots of things in there. There's lots of, I think, edge cases, you know, that it doesn't necessarily take into account.
Speaker 2:Like, what if I have this, just endless stream of AI music and it's like all bad and I'm listening for hours and hours and hours and then I find that one good thing and I snap it and then now that's mine. And like, did I did I create that or did I not create that? I I don't I don't I don't know. I think I think you'll see this change a lot in the future. The thing to remember is that copyright was these laws were invented to promote creativity and you want to, I think, very strongly enforce that.
Speaker 2:You want to make the laws that are going to incentivize people to be creative and to do new things. I think everything should be downstream of that. I would hope that we find ways of of rewarding people for doing new and creative things with art in general, not just music. And AI seems like a very obvious part of the future of those things. It's not the only part of the future, but it seems like a very obvious part of the future.
Speaker 1:One thing I've heard you say is that AI is overrated. What did you mean? What do you mean when you say it's like overrated or overhyped?
Speaker 2:It's it's definitely overhyped. I think like the number, the amount of ink spilled on on AI seems incommensurate with with, where the technology is now. Maybe that's okay, maybe that's a good thing, like you know it's it's good that people care about what I'm doing. I don't even know if the number of dollars being invested in it is incorrect, but there's definitely a lot of hype out there and that goes back to what we were talking about before where it's actually helpful company building to be a little bit removed from some of that hype, you know those hype cycles can really whip you around.
Speaker 1:As a practitioner there are things that are happening you feel like are under hyped that are happening in your world?
Speaker 2:There are things that are under hyped, that's a great question. One pattern is when you hear the same buzzwords thrown around then you never get one level deeper of detail. So, you know, maybe the latest one is agentic behavior or agents, right? Like, and then when you hear people throw it out and they have no idea what it means, this is typically a hype cycle.
Speaker 1:How do you stay balanced when all of that energy is rushing around you? And I would imagine, you know, I mean, raise a monster round, right? Like there are plenty of people who missed out on that who would be interested in investing now. How do you kind of steady yourself in such a like crazy moment?
Speaker 2:You kinda can't help but steady yourself. Like there's so much happening and you just have to focus on the things that are gonna matter. And it's like the whole thing is just an exercise in good prioritization and I don't think I'm particularly good at it, but I am, I am improving.
Speaker 1:You're figuring it out.
Speaker 2:I'm figuring it out, I am improving. And the way in which you prioritize, in an eight person company doing one set of things is very different from how you prioritize in a 60 person company doing a different set of
Speaker 1:How is that feeling now?
Speaker 2:You're in it right, so it's like very hard to describe but I, again, I wouldn't trade my job for anyone in the world, I have such a cool job, I get to work with great people on music touching you know millions of people's lives and making them smile.
Speaker 1:The shorthand in math for most people who are raising is like, wanna give myself twenty, you know, eighteen to twenty four months. I wanna have like, it seems to be rooted in something like very practical in terms of like how much time they need, how many hires they need to make. How did you think about it?
Speaker 2:It's really funny because I think maybe the dirty secret is you know I'm raising x at y post or whatever it is and then you back your way into x, you come up with x and then you come up with all of the data to justify it instead of saying that. Who knows like if this eighteen or twenty four month milestone based cycle is the right one. This is probably Is
Speaker 1:that how you thought about it?
Speaker 2:No, not really.
Speaker 1:How did you think about it?
Speaker 2:You know, I think we're fortunate enough to, I think our raising history is not like particularly common, know, we weren't running processes or anything like that. I think it
Speaker 1:Has it always been like that? Like has it always been inbound that you were fielding versus
Speaker 2:The seed round wasn't, The seed round was special because SVB collapsed literally, concurrent with that. That was that was kind of a crazy time, know, in its own Oh,
Speaker 1:insane. Totally.
Speaker 2:Very fortunate we didn't have that much money in the bank, when when that happened and we mostly had
Speaker 1:But now you have a lot of money in the bank.
Speaker 2:Yeah, maybe a lot of money in the bank, more than the FDIC limit. On the one hand, it is unclear if like the micro dosing funding is the correct way to do things, maybe it keeps you more honest, it keeps you more goal oriented, Maybe it shuts off other avenues.
Speaker 1:Here's what I love about it. Entrepreneurship is such a personal experience and you certainly can learn lessons from other people and you can take advice. You could do all that stuff. But at the end of the day, like you've really got to tune into what's right for you. And clearly like there's some degree of confidence, there's some degree of understanding of your business that allows you to step into a round like that with like real conviction versus like just pure, like totally opportunistic?
Speaker 2:I think, look, there was certainly was some opportunism in it but I think that one, you know, I say this about a lot things, there wasn't like a master plan, we're gonna do x then we're gonna do y and then we're gonna raise a big round, that wasn't exactly how it went. Two is you gain confidence as things continue to work but you feel more confident actually with the revenues because you know that there's some kind of a real business here. Maybe it's not ambitious enough to say I need revenues in order to raise this big round like maybe I should be more risk tolerant or something like that.
Speaker 1:No, don't know. There's a whole scene in Silicon Valley about that.
Speaker 2:Yeah, it's a personal thing, it's a personal thing, So I think that there's so much signaling value embedded in all of these things and it probably has people wasting a lot of time. I will reject any notion of like we've done it right. Like time will tell whether or not we've done it Yeah. Let me say one more thing then.
Speaker 1:Yeah, please.
Speaker 2:I think that here is a downside and and it is very hard when you have a lot of money in the bank for any company, I've talked to other founders who have raised a lot, other founders who have raised a little, it definitely changes your attitude toward things and toward toward spend in general, and again that is a double edged sword, right, and I think that there are reasons, like there are reasons that people kind of do the micro dosing thing that you said and it does keep you a little bit more honest, and I think just part of the having revenues is a way to keep yourself honest in some sense.
Speaker 1:How has it affected you knowing you've
Speaker 2:got nothing to with Try not to actively try not to think about it. Okay, why actively? You simultaneously have to think like what are my goals, what product do I need to build, what do I need to give people, and then you think okay, I have a war chest, how can I deploy that capital against those goals, but if you constantly have that number in your mind, it will distort how you think about things and you won't allocate it correctly? We should always be trying to think like you know, basically how do you do more with less and then that will make the capital go further so that you can do something else.
Speaker 1:Isn't that wild to say though? I have 9 figures in the bank, so how do I figure out how to do more with less? Most people would just say, hey, I've got this money hose, like let's go solve every problem with the money. As an example,
Speaker 2:I tell everyone like this is a super high trust environment. Everybody, like everybody at this company has good judgment, is very trustworthy. One thing we don't give everybody is Stripe access because we don't want everybody constantly focused on the revenue numbers. And if you ask me, I'll tell you what they are. If you if you are constantly focused on that number, it distorts how you think about things.
Speaker 1:Is that a lesson you learned yourself? Were you Did you ever focus on that too much? Yes.
Speaker 2:Okay. Yes, and in fact, yes, yes, a %.
Speaker 1:Okay, say more. What did it do to you?
Speaker 2:You are constantly, you know, like when you're going vertical and you feel really good and you're just constantly refreshing that page, it's this is like the version of social media of like you got addicted to the thing and you're not actually doing anything productive and so like let's stop it.
Speaker 1:How'd you break that habit?
Speaker 2:I don't know, discipline? Or who says I have but, Have you? Yes, but, but, discipline, discipline. It's just like focus on the things that are going to matter because if you don't, no one else is. You know, another great piece of advice that I got from an investor is like, doesn't matter what it is, it's always the CEO as well.
Speaker 1:It's true.
Speaker 2:Always the CEO as well.
Speaker 1:Always the CEO as well. How or whether you feel the pressure to spend it because it's there and how do you kind of manage that psychology? Because I would imagine the starting gun's gone off, I took on all this capital and now I'm gonna, you know, tripling the team and I'm building all this stuff.
Speaker 2:You should feel you should feel pressure to use the money correctly and efficiently and quickly and to I'm not sure the cost benefit calculation of is x worth it. Should that be dependent on how much money is actually in the bank? It's like, is x worth those dollars? Yes or no? I actively am trying to avoid having this company turn into like, just a great source of waste because there's a lot of money in the bank.
Speaker 1:Have you ever felt out of your depth?
Speaker 2:What every day. What do mean? Who the heck? I'm making this up. But how do you feel?
Speaker 2:What what what business do I have starting a consumer music company?
Speaker 1:Right.
Speaker 2:I know a little bit of music and that's it. I have no business doing this.
Speaker 1:Yeah. Why are you doing it then? Because it's fun. Or how are you doing it then?
Speaker 2:You know, like what are my strengths? I enjoy interviewing people, and I actively enjoy not being the smartest person in the room. Yeah. Most of the good ideas here are not mine. And, and so you find great people and the rest takes care of themselves.
Speaker 1:Have there been any, like, truly transformational ideas, like someone came up with that just completely sent you off in a different and more productive direction?
Speaker 2:Yeah. Tons. Like, you know, we started on Discord and I thought we'd be on Discord forever because I look at mid journey and they're just like, they're crushing it on Discord. And Martin, one of my co founders is like, this is so stupid. Like, we could build a much, like, more better, a much better, more intuitive user interface.
Speaker 2:I'm like, no. No. No. No. No.
Speaker 2:No. And then, like, for whatever reason, we release a very thin web app that is just like a strict subset of what you can do on Discord, and it takes five days and 90% of the traffic is on the web.
Speaker 1:Shut up.
Speaker 2:I'm just like Five days? I couldn't be more wrong here. Right? Like, yeah, five days. This was this was just wild.
Speaker 2:So many things I got wrong to that degree. That that is like really just entirely wrong. Right?
Speaker 1:How do you find that humility though to Same
Speaker 2:with mobile also.
Speaker 1:Okay. So, say more about Like,
Speaker 2:we need a mobile app. I don't know, people seem to like the web app. We need a mobile app. Right? Yeah, mobile is obviously the right idea here.
Speaker 2:Just be open minded.
Speaker 1:You're very good at deflecting your own contribution here, but when you sit back and look at Suno, what do you like where's your fingerprints on it?
Speaker 2:Just the people. Just the people. Really? Yeah. It's funny.
Speaker 2:It's funny. Like you're not gonna not gonna believe me, but, like, I'll even, like, point out when an idea is good and mine because it happens so infrequently. One happened to happen to happen last week.
Speaker 1:I'm terrible.
Speaker 2:I can't tell you.
Speaker 1:But Come on. Drive me crazy.
Speaker 2:I can't tell you.
Speaker 1:It's not
Speaker 2:that good. But but, no. I'll, like, literally text someone like, I have a good idea. You know? If I have my rare like, you know, I have my I have my biannual good idea.
Speaker 2:It's just the people. It's just the people.
Speaker 1:I love it. Well, this is fun.
Speaker 2:Thank you.
Speaker 1:Thank you for making time.
Speaker 2:Thank you. Thank you for coming.
Speaker 1:I hope there were some things in here that really connected for you. We'd love to hear about them in the comments or in a reply to the newsletter. If you're working on something that fits the profile of Suno or other things that are seeing small teams having a outsized impact on their market, This is an area that we're incredibly interested in for investment. If there's something here that resonates with you and how you're building similarly, that you'd reach out to hellowindi.vc. Take care.