Intro
CEO of Suno Mikey Shulman went on Training Data hosted by Sonya Huang, the podcasting arm of Sequoia Capital (which manages $56B in assets), to talk about how everyone can make music now thanks to their incredible product. This image alone sent me into a writing spiral, what you see is the culmination of that work. I'll discuss in this article how Silicon Valley doesn't understand music or artistic creation, and how framing music in this way destroys any actual novelty or joy. Sorry for the length.
Suno, Scourge of Singers and Songsters
Suno, for anyone unaware, is an "AI" music generation platform. According to them however, they're "a music company built to amplify imagination", which is so obtuse it borders on a lie. Firstly, music companies don't exist. There's labels which pay for and manage music, there's artist compounds and communes where people work and create together, and there's instrument makers that craft the thing we can use to make specific and interesting sounds. There's concert venues that host music, companies that make speakers, headphones, and microphones, even schools where you can learn how to make music and refine talents. There are not music companies though, as music is a thing created and experienced.
Instead, Suno resembles a music factory, pumping out whatever is prompted. It's unlike other factories however, as much of its material was stolen, there's no quality control, and lacks any of the process and refining that makes music music. Put another way, it's a 3D printer that you can prompt for music, but it has no concept of what "music" is, and none of the refinement that goes into making music beyond "well, I can technically hear this with my ears." Also, all the filament for it is stolen, somehow.
One key part of understanding where Suno, and Mikey by extension, comes from is to take a look at their company values. They are as follows:
- Music is our company focus. We're a music company built around a single purpose: transforming how people create and experience music. AI is our tool, not our identity.
- Impatience is a virtue. We always strive to improve our bar. Extreme speed helps us achieve that.
- Aesthetics matter. Great taste isn't subjective—it's a skill. We trust our instincts, value beauty in every detail, and have the courage to follow good judgment even when the data disagrees.
- Fun is underrated. We take joy seriously. Creativity thrives when we're playful, open, and having fun—because making music should feel as good as it sounds.
Music being their focus comes as no surprise. If you're going to run a company that shits out "music", I would hope you're at least focused on it. I can even see the argument for them changing how people create music, even if I think it's detrimental to art as a whole. However, I can't see how this changes how people experience music. It's a buzzword that plays nice with the idea that they're revolutionizing all of music. As I'll come back to many times, sounds nice but doesn't really mean anything.
Impatience as a virtue really irks me. Silicon Valley already has its notorious slogan of "move fast and break things". To value constant change is a loser's perspective to begin with, but being impatient about that change speaks to a deeply sick culture in the company. Ideas and projects are honed through time spent sitting with it, reflecting and considering why and what we're doing. Speed is not an indicator of something being good - and it never has been. I can't see a world in which this mindset creates anything other than a sprawling, mismanaged project, and I hate to think that that is what the "future of music" is supposed to be according to them.
Aesthetics matter only insofar as this product allows a user to create something with the veneer of skill and craft without any of the knowledge or effort required to do so. Taste is a skill, yes! Suno hones it in the same way an infinite amount of monkeys bashing at typewriters eventually improve their writing skills though. Applying even a marginal amount of effort to actually creating something with your hands is infinitely more valuable in developing a sense of personal style and taste. The output from Suno can be as polished as you want¹ and you'll never have to think about nor work for the end product. Simply push the music button and music comes out! No, it won't mean anything to anyone, nor will it have any thought behind it, but you'll be able to listen to it! Isn't that fun?
Fun in this context feels like it speaks to a person that I can't understand and don't want to. This is a person who either a.) doesn't respect the artists they like enough to try and understand and see the time and effort that went into making their art, or b.) doesn't respect themselves enough to try and express their own genuine perspective or experience. There is fun to be had in both parts there. Instead, Suno tries to manufacture a person whose enjoyment is so removed from the creation and creator of the art that they'd be just as happy pushing the Music Pellets button until it breaks. Or, a creator that is so disconnected from their own work that it doesn't bother them that part of them is reflected in this "art" you just "made"?²
We return to Mikey, king of Suno the music printer. We have a better idea of his company and its goals. So what does he have to say about his product?
Making Sense of Music
Originally, we thought this would actually be too hard. And it’s because you have to rewind. This is pre the ChatGPT moment. We did some back-of-the-envelope math. We knew we loved audio, but the back of the envelope math told us that actually producing good music, making good music, generating good music was a couple of orders of magnitude away in terms of compute and model size and capability. And it’s because music, sound in general, is very unwieldy. It’s not in discrete bits like text is. And so we actually started building a company that was all around using the same technologies to make sense of audio, not to produce it. And very happily, pretty early on, we had the right breakthroughs and we realized, oh, we actually can make music.
-Mikey Shulman on the start of Suno, emphasis my own
This type of statement is common across AI industries: "Oh, once the bot understood the prompt...", "It just feels like Claude gets me", and "Suno can make sense of audio...". It shows me who Mikey is; someone who does not care enough about what he's doing to speak accurately, but rather keep it tight, snappy, and simple for any moron that might be listening. He did the same thing at the very beginning of this podcast, saying "in Western music, there are 12 tones"³, which sounds deep until you think for a moment. In reality, it's a vastly oversimplified way of framing things that goes so far as to be misleading when he's the one speaking as an authority here.
These pieces of software understand nothing. Suno does not, in any sense of the word, make sense of audio or music. It cannot by the very nature of how the programming works. I'm not going to go into a long, technical description of how LLM at large work, if you're interested I've found this visualizer of LLMs to be the most approachable. While audio is certainly dissected and weighted differently, we have no reason to believe, nor any indication from Suno, that their technology works differently. I'm sure they would be shouting it from the rooftops if it did too, as that would be a novel structure for LLMs and likely massively profitable to license out.
Suno doesn't understand dick about music, nor what you prompt it with. One of the most advanced tools being paraded around, ChatGPT, described a series of fart sounds as "atmospheric" and had a "nice bedroom/DIY texture". Suno is simply a very complicated paste dispenser. It cannot and will not understand nor ascribe meaning to anything it does. No choices were made, there was nothing subversive in an output. It is no more insightful in its output than a calculator doing simple math. We just love to project feelings onto things that we think act like us. People get attached to rocks with googly eyes, and machines that just parrot things back at them.
So again, Mikey does not care about being honest about his product or is too high on his own supply to worry about how it works. Instead, he points to his magic box and insists that it's just as good as making music yourself, while quitely implying that it's just too frustrating and boring to make music the old way. You wouldn't want to do things slow or be bored would you? You've got content to make!
Culture & The Lack Thereof
About halfway through the podcast, Sonya asks about the cultural and memetic tendency of music to be shared, and how AI music plays into that. To his discredit, Mikey largely dodges the question. He agrees entirely that music by its nature is largely a shared thing, that humans love to perform and attend concerts and make mixtapes. He fizzles out at the end though, saying "...the thing that’s amazing is that AI can be used to actually change that and to augment how music is perceived in society and in culture, augment how it is used socially, because it’s actually become less social in the last 30 years."
No evidence to his claim that things have gotten less social, just vibes. Really there's counter-evidence, as over the last several years live music has been on a noted uptick, there's more ways than ever to share music across the various social media feeds or dedicated apps, and people can still make playlists, mixtapes, CD's and whatever else. It's just a convenient narrative to help sell Suno as an experience and as a platform.
In reality, any social aspect of Suno is not linked to it being AI, its just basic discoverability; the same you would find across platforms like Apple Music, Spotify, and Bandcamp. The difference here obviously being that anything shared on Suno is definitely slop garbage, instead of there just being a chance of that on the others. AI music and AI tools do nothing for the social aspect of music, the thing that brings people together to bond and make memories with it. I'd argue that creating on Suno is actually less social, as you don't need to interact with others to get different sounds or instrumental expertise, a few prompts and you're on your way. What Suno does enable is the flooding of other platforms with even more low quality crap, harder than ever to filter out.
Tediously Breaking Down Nonsense
Mikey lives in his own world. I don't say that to mean he's unwell, but has such wealth and power enough to bend reality around him. His experience is completely disconnected from the reality a majority of people experience, and actively contradicts itself because of it. His life is so optimized and focused on this singular goal it tries to rewrite human existence. That sounds like a lot to take from a single response in a podcast, I understand, but I stand by it. What follows is his response in its entirety.
This is the crazy thing about Suno. Before Suno, basically everybody was a consumer of music. Compared to the eight billion people on the planet, there are very few people who make music and the rest of us consume it. And that’s fine. It tends to cater to passivity. It tends to cater to making it less social and more impersonal. And the crazy thing about Suno is that on any given day, 90 percent of the users are going to create something. And the thing that’s hard to wrap your head around is you’re not creating it to go bring it elsewhere, by and large, to do something with it. People are creating music for the fun and enjoyment and fulfillment that comes with being creative. And so that, the creation, is actually the entertaining bit. And that is the big step change. It’s that everybody in the world is creative. Being creative makes you feel a certain way; this is in our DNA. And we are basically using technology to allow everybody to feel those warm and fuzzy feelings. A lot of the inspiration for me personally for doing this comes from just remembering the fondest memories that I’ve had, or some of the fondest memories that I’ve ever had are making music with my friends. Not even performing in bands but, like, practice was so much fun, and you get really close to people making music. And it’s because it feels really good to be productive in a way that doomscrolling your favorite app for an hour does not feel so good when you’re done.
-Mikey Shulman, replying to the question "Are people more creators of music or are they more consumers of music or both?" during the podcast
This statement is so incredibly dense with contradictions and misunderstandings that I'm just going to go nearly line by line. As promised, here is a tedious breakdown of Silicon Valley nonsense:
- Before Suno, basically everybody was a consumer of music. Compared to the eight billion people on the planet, there are very few people who make music and the rest of us consume it. And that’s fine: This attempts to posit a world where there were a select few who were skilled and had access to the right tools to make music before Suno came along. It's a heavily revisionist idea of the history of music, given that a.) everyone has access to tools to make music, right now! You have hands and a voice or a way to rhythmically smack something, whether it be log or bongo or drum pad, and b.) all cultures since time immemorial have a rich history of making music, whether it be hymns, folk, work songs, the list goes on! People make music with what's around them. If you don't, it's not for lack of options, it's for lack of effort. If you can't attain a sound you want or a note, that's a different issue, but it's not one where people can't make music.
- It tends to cater to passivity. It tends to cater to making it less social and more impersonal: This is just a convenient narrative to try and drive home that they're shaking things up and democratizing music. People share music all the time, go to concerts all the time. Making music can be social, whether you work with a group or just jam at a friend. Any passivity is purely by design by the streaming platforms or for lack of trying on the person's part. It simply isn't a real issue going on.
- And the crazy thing about Suno is that on any given day, 90 percent of the users are going to create something: So, this statement is a bit of verbal sleight-of-hand, as it's pretty obvious on its face but phrased like something novel. All he's saying is that 90% of the people that end up on Suno make something, which is probably hyperbole but I'll go with it. If you're ending up on the site for music generation, yeah dog, you're probably going to generate some music. If you end up on Twitter, most users will look at their feed. The purpose of the site is what people do on it, no shit.
- And the thing that’s hard to wrap your head around is you’re not creating it to go bring it elsewhere, by and large, to do something with it. People are creating music for the fun and enjoyment and fulfillment that comes with being creative. And so that, the creation, is actually the entertaining bit: I guess I'm supposed to be impressed by people who end up at the music generation site generating music when it's as simple as typing a few words and waiting. When the barrier of "creation" is so low, the amount of enjoyment someone has to expect to seek it out is low too. It's the reason he finds actual practice and play so annoying, he wants to skip the parts that give the joy to the act. All reward with no effort. I'm sure the users would feel more fulfilled if they actually took the time to learn and create something too, but they're not going to because they have the button that gives them morphine until they die right here.
- And that is the big step change. It’s that everybody in the world is creative. Being creative makes you feel a certain way; this is in our DNA. And we are basically using technology to allow everybody to feel those warm and fuzzy feelings: This one makes me feel crazy. It doesn't take a fucking world class pianist to type "make a rock song with a sitar and an MGK feature". And as previously stated, everyone already could stomp their feet, tap rhythmically, whatever. If your head isn't full of rocks, regardless of what we call the creative part here, no one should be shocked that everyone can do this. If we a machine that takes in that prompt and spits out something that even passably like music, no duh people are going to get warm fuzzies like they did something. You just spoonfed them dopamine with a musical wishing box.
- A lot of the inspiration for me personally for doing this comes from just remembering the fondest memories that I’ve had, or some of the fondest memories that I’ve ever had are making music with my friends. Not even performing in bands but, like, practice was so much fun, and you get really close to people making music: I don't see how we get to Suno when some of your favorite memories are making music with your friends. They are just such fundamentally different things, I think he's just twisting his own past to support where he's at. Like yeah, jamming with friends is incredible, Suno replicates none of that, and at best it would be just that: a replica. At worst, it's a soulless husk that profits off of the work of nearly all recorded work to depersonalize it and profit off of it.
- And it’s because it feels really good to be productive in a way that doomscrolling your favorite app for an hour does not feel so good when you’re done: Suno has this weird attitude that's hard to pin down. On the one hand, he speaks of music as being such an incredible thing to just experience, but on the other it's about productivity and output. Things need to be perfectly polished and tweaked just so. There's a subtext to all of this that learning, trying, and failing is just unnecessary and embarrassing, that we've moved past that need. In reality, you cannot produce anything of value without that effort, without trying and failing.
To Slop, Or Not To Slop
Following this answer, they talk about how Suno is actually a creator platform, not trying to replace Spotify. Sonya says "...You're turning people into creators" and Mikey replies "A hundred percent." Creator is distinct label from artist; one a lot of artists bristle at, and why so many hate having to play the social media game to get a crumb of attention nowadays. Yes, people can be both a creator and an artist, but a creator has a vastly different perspective on what they make than an artist for most folks. It has a different implied set of values and interests at play; one is likely to be much more comfortable with pumping out slop than the other.
A softball is lobbed after in the form of what Mikey thinks about the label "slop", and I just hate how he sees the world. He tries to equivocate slop with anything a majority of people don't/wouldn't care about, which is just not what it is in the larger cultural use. A song he makes with his 5 year old is not slop. Is it something ready for posting to the whole world? Maybe not, but that's okay! I bet a lot of people would honestly like to see that (if it wasn't made with AI, like I'm sure it was). I mean hell, I follow a guy who just makes songs from lyrics his 3 year old daughter writes, that shit is so fun!
What makes some output slop and some output not is effort, and he knows that. The less effort you put into something, the more slop-like it is. He tries to muddy the waters by making it a black and white issue; slop or not for him is determined by quality and polish. Hence why he brings up 13 year olds making beats off of laptops, albeit it doesn't make the point he is hoping it does. Plenty of huge artists got their start making beats off of pirated FL Studio in their bedroom. Were they all incredible? Surely not, but no one is going to call it slop, because it was something that they put heart and soul into, trying to expresss something. It's just fucking dumb: art is not lesser because it's less presentable or polished.
It is less if you intentionally take shortcuts and get mad at people who don't appreciate that. There's a distinct difference between working smarter and undermining what people do, making bad copies. Typing "country song with a trap hi-hat and a Justin Beiber feature" does not imbue the associated output with any sense of expression or effort. "Sad song with harps and a soulful piano" cannot reveal the deep, aching of sadness in an artist from a life of neglect and mistrust like an actual song does. What is achieved is a simulacra of music; a hollow shell painted to resemble it. What is generated by Suno is categorically not expression. Some people will express themselves over it, sure! But nothing born from Suno can ever or will ever express something on its own. It is content rather than art; it's caring more about an end product than the process. Suno and its ilk don't get to be mad because people treat it different.
Mikey then mentions successes that have come from Suno, people getting signed record deals or chart topping tracks, but that doesn't undo the slop of it all. That just means that some people enjoyed it and/or thought they could make money off of it which is up to them at the end of the day. While I think they're wrong to do so, they're not the ones making it, and the people listening may have been genuinely unaware what they were listened to was generated. Throwing in that he "just know[s] that there are tons of charting tracks that have little bits of Suno in them" feels like trying to build a narrative of everybody already being on board. All the big artists are doing it, why not you indie artist? Don't you want chart toppers with that sap the soul out of you? Just trust us, it's for the best!
It becomes borderline comical when he goes out of his way to say "actually, record labels are pretty cool!" (paraphrased). I would expect him, like all the other AI outfits, to be talking about a "democratizing" effect, how it gives artists tools that the labels keep to themselves or some shit (which would be untrue). He knows where his bread is buttered though. If there is to be a long-term plan for making AI generated music a part of the music industry, the labels will be involved. What he describes as "building products that let fans interact with their favorite artist and really deepen the artist-fan connection" sounds like an absolute fucking nightmare for a human artist, but extremely profitable and a non-issue if the artist doesn't really exist.
One Last Note About Capitalism
As they wrap up their conversation for the podcast, they joke:
Sonya Huang: Yeah. And what are we going to do with all our time after all the robots are doing all our work? Like, we …
Mikey Shulman: You’re not going to want to doomscroll for an hour. You’re going to want to be productive and fulfilled.
It's this convenient lie the people with money weave, intentionally or otherwise. For them, it's true. At some point, AI will be able to do everything that makes them money and more. They won't have to worry about a thing; money in, more money out. But it can never and will never work like this for the vast, vast majority of people. The ruling class will be happy to displace anyone they can with these LLMs. This just goes to show that even the arts aren't safe, they'd be more than happy to replace the creatives as well.
These programs won't be a force for good, not unless they're forced to be. The markets won't make that happen as it's much more profitable to bleed people dry. It will take action from people and their governments for there to be any semblance of equity as these tools continue to expand their influence.
What Are We Doing Here Mikey?
I feel bad for Mikey Shulman. I mean, I hate what he's doing and he's fundamentally trying to undermine the artistic process, but at the end of the day I have to feel bad for him. He demonstrates a severe case of Silicon Valley brain with AI psychosis features, which is particularly hard to cure. Somewhere in his history, he came to completely misunderstand art and the world around him. Acting like his company has suddenly unlocked a creativity latent in people that they could never access before! It's a deeply misguided and disrespectful view of the world, centering him and what he thinks is important about music, which is purely the product.
Art is more than the constituent parts put in. Through the process of shaping, thinking, and reshaping, the material is imbued with something a little extra. A bit of the person creating is put in everything we make. Suno attempts to subvert that process and instead jump straight to the end product. In that action, we lose what makes art art. I hope Mikey comes to his senses and burns down Suno in a fiery display of artistic reclamation. Failing that, I hope he can come to terms with how much he's perverted the act of creation in pursuit of investor dollars. I'll be making art the old-fashioned way in the meantime.
Thank you for reading. Here's a homegrown meme for your troubles.
Footnotes
¹ It can only be so polished because it has hoovered up the sum total of recorded human expression, which was largely stolen or licensed for pennies on the dollar. Their facsimile can be so convincing because it's been able to absorb millions of datapoints from every genre, without any of the homage, thought, or extrapolation that actual artists do when drawing on inspiration. [Return to Article]
² I don't have enough to flesh this out into a full point, but after talking with my wife I think a lot of the appeal for users in Suno comes from a hustle culture mindset of "Get stood up quick and make money, it won't be here long". If someone respects music, the people that make it, and its history, I don't think they're taking the "shortcut" of Suno. However, if you see music as a side hustle or "hobby" to post on socials and grow the brand, it's gotta look good! You don't want to post subpar content. So you use Suno to make it polished and "good", despite being fundamentally unable to actually make what you've claimed to make. It's a Shopify template selling AliExpress crap pass at music.
To run it back for a moment, I think that's all it is for Suno users, content. For those doing it for the love of the game, they see music as not worth paying someone or making yourself, just wanting to be catered to like a little prince and move on. Absorb and throw away. Even if they create something they think they love, what then? What's the end goal if not just customized content for any moment they wish? [Return to Article]
³ To be clear, he seems to be gesturing vaguely to the chromatic scale, although it's hard to tell since he just says "In Western Music there are 12 tones" with no other context or reason. That scale is important, sure, but is a thing that humans made up for writing and composing music, and does not by any means contain all of music, Western or otherwise. I also take issue with specifically talking about Western music here, which feels like a blend between a dogwhistle and an attempt to sound more educated by reciting something he learned in a music theory 101 class or on YouTube. It's gold leaf on a Ritz cracker; trying to dress up a wholly unimpressive and frankly irrelevant statement. [Return to Article]