Hacked - How to Brute Force Music

Episode Date: September 15, 2023

The story of two guys who set out to put every musical melody in the world into the public domain, and the technology they hacked together to do it. Featuring Damien Riehl from AlltheMusic.info. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Heads up. There's some light cursing in this one. Think of a musical melody like a sequence of numbers. If you say the first note of a scale is one, and each subsequent note in the scale is two, three, four, five, six, seven. Melodies are just sequences of numbers. Happy birthday to you is just one, one, two, one, four, three. And if there's one thing that computers are really, really, really, good at, it is numbers. And one evening, after a long day at work, Damien Reel and his colleague

Noah Rubin have this idea. After a 14-hour day, my colleague and I were having a beer at the lounge at the hotel, and I said to Noah, I said, hey, Noah, you know how we can brute force passwords by going A-A-A-A-B-A-C? I said, what if we could do that with music by going do-da-da-da-da-da-da-da-da-da-da-da-da-da-da-da-da-a, don't know to me, do nother to fa, until we mathematically exhaust every melody that's ever been and ever can be. And he said, fuck yeah, let's do that. And so fuck yeah, they did that. They started on this project, not to create a program to write a lot of melodies, but a program to write at 300,000 melodies per second, all of them. Every melody that ever has been or ever will be. and then to do something very unexpected with all of those melodies that they generated.

But we'll get to that. First, we got to talk about the Beatles, specifically George Harrison. In 1970, after the Beatles breakup, George puts out what is going to be his first big post-Beatles hit, My Sweet Lord. And a few months later, the record label of a totally different band, The Chaffons, reaches out to George to let him know that they are going to be suing him. They allege that his song infringes on the copyright of a song by the Chaffons called He's So Fine. And the suit was explicitly concerned with the melody. And the conclusion to this case is really important.

The court basically decided that while George Harrison might not have consciously infringed on their song, he might have subconsciously infringed it, meaning he heard the song somewhere, it got stuck in his head, and then he unknowingly wrote a melody that copied another one. But the problem with that is that that essentially assumes that every defendant is guilty, that is infringing. because how can George Harrison prove that he's never heard a song before? Philosophically, it's impossible to prove a negative. So he can never prove that he's never heard it on the supermarket loudspeaker

or on a friend's phone that he held up to it, right? There's no way to philosophically prove a negative. So every one of these defendants has lost after that case because you can't prove it. All the plaintiff has to say is, hey, my thing has been seen three million times, so you must have seen it. You must have done it. So this is the backdrop upon which I said, this is really stupid. So he and Noah did something about it.

They created this program and told it, just start writing melodies, brute forcing them like a password. And a few million and eventually a few billion melodies later, rather than just sitting on them. Or worse, suing people saying, ha, we own these now. They took all those melodies they'd written. And they placed them in the public.

domain, freeing them. Because from that point on, anybody who gets sued over a melody can say, well, I wasn't inspired by your song. I was inspired by a melody by Damien and Noah. And their stuff, well, it's in the public domain. The questions raised by Damien and Noah's project, called fittingly, all the music, have never mattered more than they do in a world where generative AI has empowered people and corporations to generate huge volumes of stuff that could reasonably have been authored by a human. What it means to have authored something, to have created it, to own it, it's muddier than it's ever been.

So I called up Damien to talk about all of it. My conversation with lawyer, musician, and technologist Damien Real from all the music here on Hacked. Thanks for taking the time to talk with me, Damien. I really appreciate it. Hey, thank you. I really appreciate you. Ask me out. So just to dive right in, this project, all the music, I understand that it's been brought up in musical copyright cases. So to start, generally speaking, what are these You Stole My Melody lawsuits and which have this project come up in? Sure. So the You Stole My Melody lawsuits are largely things like Katie Perry got sued

for a melody by a guy named Flame. Flame said that Flames' melody was literally this. Dun, dun, dun, dun, dun, dun, dun, just quarter notes, descending from the tonic down to the seven, down to the six, right? So really, that dead simple thing, a jury decided that that was $2.6 million for infringement against Katie Perry. So that's the backdrop of insanity that I was against. I gave a TED talk on this, that that Katie Perry verdict came out on a Tuesday, and my TED Talk came out, it was done on Saturday. So I integrated that into my TED Talk, where my project, which we're going to talk about in a bit, is a way to brute force melodies. And that particular melody from the Katie Perry case that showed up in my dataset 8,128 times. And it was

created at the rate of 300,000 melodies per second from my machine. So the question is, should Flame get a monopoly, and really what copyright is is a monopoly, should Flame get a monopoly on for life, his life, plus 70 years monopoly for something that my machine shit out at a millisecond. And I don't know if I can swear on your podcast, but... Oh, you can. All right, good, good. So the question is probably not, right? Because it's not, his is not sufficiently original to be able to say that you should be

able to do it, because to be copyrightable, it needs to be original. So now let's go back to what is my project in the first place. When I was working at, one of the things I was doing is Facebook, I did cybersecurity for a while. Facebook hired me in my company to investigate Cambridge Analytica. So I spent a year of my life on Facebook's campus with Facebook's data scientists. And my former FBI, CIA, NSA people worked with me figuring how bad guys use Facebook data. After a 14-hour day, my colleague and I were having a beer at the lounge at the hotel, and I said to Noah, I said, hey, Noah, you know how we can brute force passwords by going AAA, A-A-A-B, A-A-A-C-A-C-A-C-A-C-A-C-A-C-E.

I said, what if we could do that with music by going do-da-do-do-do-do-do-do-do-do-to-me, until we mathematically exhaust every melody that's ever been and ever can be? And he said, fuck yeah, let's do that. So that night, we did a prototype of about 3,000 melodies that he did with a prototype. And to date, we've done now 471 billion with the B melodies, arguably exhausting every melody that's ever been and every melody that can be. So the flame melody I talked about earlier, that done-d-d-d-dun-dun-d-d-dun. that showed up at 8,000 times in my 68 billion, melody, 68 billion data set.

So all that's to say is that all of these cases before mine have lost. That is, you know, everybody knows about the blurred lines case, and everybody knows about all of the other cases where George Harrison is the most famous, where he, the court said, you know, George Harrison, I don't think you consciously infringed somebody, but what I think happened was you subconsciously infringed them. that is you heard the song somewhere and it's stuck in your brain and then you wrote the song that copied the other song. But the problem with that is that that essentially assumes

that every defendant is guilty, that is infringing. Because how can George Harrison prove that he's never heard a song before? Philosophically, it's impossible to prove a negative. So he can never prove that he's never heard it on the supermarket loudspeaker or on a friend's phone that he held up to it, right? There's no way to philosophically prove a negative. So every one of these defendants has lost after that case because you can't prove it. All the plaintiff has to say is, hey, my thing has been seen three million times. So you must have seen it. You must have done it.

So this is the backdrop upon which I said, this is really stupid. So my argument is that maybe those melodies, like Flames melodies, are so common and so unoriginal as to be uncopyrightable. Therefore, those shouldn't go. So I made that argument in his head talk. And to date, there are several versions of the TED Talk that are going around. They've all collectively been seen about 2.1 million times. And my whole goal with that is to have judges and the judge's clerks and the lawyers look at that and say, hey, maybe this is a good argument that may be unoriginal, therefore uncopyrightable.

So every defendant before my talk lost, every defendant after my talk has used that argument, that unoriginal, therefore uncopyrightable, and has one. They have not cited me. And so correlation is not causation, but there's a really good correlation that afterward every defendant, including Katie Perry. So Katie Perry, remember, got dinged for $2.6 million. My TED talk came out then. And then after that, the judge reversed the jury verdict.

That for the thing saying that that melody as a matter of law is unoriginal, therefore uncopyrightable. The same argument that I made in my TED talk. Same thing for Led Zeppelin. After my talk, the court said that the stairway to heaven, unoriginal, therefore uncopyrightable. Similarly, Ed Shearin, in his UK case, unoriginal, therefore uncoperioder. So I've got a pretty good track record. Everybody before has lost.

Everybody has made the arguments unoriginal, therefore uncopyrightable, and has won after my talk. Interesting. So music contains a bunch of different elements, obviously. There's rhythm, there's melody, there's timbre, there's the arrangement of the song. A lot of these lawsuits seem to center in very specifically on the melody. And at the heart of your argument is this idea, and I don't want to put words in your mouth, but that melodies are more like mass.

Mass like any other fact can't really be owned by a person. Can you explain that idea for people? Yeah, absolutely. When you think about what is a song, a song is a component of many parts. One of the parts is a chord structure, right? One of the parts is rhythm, beats. Another part is, of course, melody. Now there's timbre, right?

Each of these are component parts of what is a song. And so the courts have said long ago that chord structures are uncopyrightable because there are only so many chords, right? If you were to copyright the blues scale, would we have blues anymore, right? Of course, or blues courts, right? You never have blues anymore because there are only so many, only so many chords. So we've said for decades, chords uncoparitable. We've also said for decades drumbeats, uncopyrightable, because there's only so many way to do a backbeat, right? or so many ways to do with a swing, right?

So you can't copyright the chords. You can't copyright drums. Really, the next step is can you then copyright in isolation the melody? And maybe in what my project is saying is maybe the answer is no, right? Maybe, you know, twinkle, twinkle little star is the same as blah blah, black sheep, have you any wool, is the same as ABC, D, E, F, G, right? Each of those is identical. But when I say that to people that are not musically literate, they're like, holy shit,

I had had no idea that those have the same melody. I said, that's because there's different damn songs, right? These are different songs that live in different spaces in their mind. Even though they share the same melody, they're different songs. And so I would say that my project is really bringing to light that, hey, maybe even though the things have the share the same melody, maybe it's not because it was copied. Maybe you independently came to the same melody. And maybe we should just all write songs and not have to worry about having targets on our

back for maybe songs that we've never heard before. You have this flow chart in the frequent. The incorrectly named frequently asked question section on your website, because that is not a fact. That is a novel. And it's really cool. But the flowchart explains how to understand whether or not you think these melodies can be copyrighted. It's essentially a fork.

Do you think they can be copyrighted, in which case you have sort of created this defense? Do you think they cannot be copyrighted, in which case all these lawsuits are sort of moot? Can you explain that fork for people? Yeah, absolutely. There's a real question these days, especially with generative AI, is if machine generated then copyrightable, right? And so the idea, of course, my 471 billion melodies are machine generated. You know, we said what are the meets and bounds?

What are the, and maybe four-year more technical people, they might say, what is your data set? And what we did is we took every, within, the first data set was within an octave, do remi me falsoleti-do. And then we took, for the first data set, I think, 12 repeated notes. So we went do-da-da-da-da-da-do. And the next one was do-d-d-da-d-d-d-re. Right, 12 repeated.

So that 8 to the 12th power is 68 billion. And then that was the diatonic scale. And then we did the black notes next. So we did the chromatic scale. So 13 notes, including the tonic on top. That is 13 to the 10th power. And I think that was about 62 billion. And then we incorporated rhythm.

So what we did is we added a note at the top that was silence. So essentially what is rhythm, but essentially a note punctuated by silence. And so then we added that. So essentially we just did a bunch of different versions of all the things, to make 471 billion of these. So anyway, these are me and Noah, Noah Rubin, my collaborator. We were setting what the parameters are of the up and across.

And then, but the machine was actually just going through all of those. So the question for the Copyright Office is, are those copyrightable if they're machine created? That is we've actually put, is it like a typewriter? Like I've essentially did the typewriter and I said, this is what we need to do? Or is it more like generative AI where you just say, make a thing and it makes a thing without any control.

So that's the real question we're thinking about if is it machine generated, therefore is a copyrightable. The US Copyright Office has said definitively, if machine generated, then uncopyrightable. So let's, well, forget what I just said. Let's pretend that it is copyrightable. If it is copyrightable, I've now copyrighted 471 billion melodies and then I put everything in the public domain to protect you stole my melody lawsuit defendant. So if it is copyrightable, Everybody's probably okay because everybody knows that now my 471 billion are in the public domain. That's at least one argument. But now let's go back to the Copyright Office saying it is uncopyraitable.

If it is now uncopyraitable, why is it uncopyrightable? And I would say the reason it's probably uncopyritable is because it's unoriginal. If on original, then uncopyritorable. So the question then is, if my 471 billion are all unoriginal, because machine created, if a human makes an identical one, melody, to any of those 470 billion, does that magically turn it from unoriginal to original? And I would say the answer is no.

And so that's really what, so either way, whether it is copyrightable, therefore public domain, therefore people are probably okay. That's one route. But the other way is even if it's not copyrightable, why? Is it because it's unoriginal? And if so, then if unoriginal,

then a human created one would similarly be unoriginal. Is there a Malay person, both technically and legally on this, no bit about music, but is there an argument that it's not a question of, because it was produced by a computer program that you and Noah made, a person could theoretically own this melody. You just don't because you didn't create it. You created a piece of software that did. Is that potentially an argument against this project being effective? Maybe it is, but I guess who would make that argument?

I guess the whole purpose of this would be in a court of law, right? And so this would be essentially what your argument is kind of shadowboxing. But, you know, I could say that, you know, I've created the melody, dun, dun, dun, dun, then, then, then, then, then, therefore that's totally original, right? But I think if you go to a court, that court would say, no, that's unoriginal because it's just a diatonic scale going down from the 8-7-6, right? So anyway, so really somebody could make that argument that I as a human did it, but it didn't work for the same. I think the reason this project kind of keeps going viral and blowing up every so often is because

everyone who hears it recognizes that you had the opportunity to do something, I don't know how else to put it, kind of evil. You had that chance the second you created this program churning out all these melodies. And yet you did it. Do you ever think about what someone could have done with this project that didn't hold the values about human creativity, that you, I guess, just happen to hold? So, yeah, so Noah and I are both do-gooders. We like to do good in the world. I contribute to a bunch of open-source projects.

He contributes to a lot of open-source projects. So when I initially did the project, I said, if somebody did this that is nefarious, this could be really bad. So I said, what we need to do is we need to make as many of these melodies as possible so that the bad guys can't be able to do it before we as the good guys are able to do it and then put it in the public domain. And we have to demonstrate this proof of concept for good before somebody starts using it for evil. And so anyway, so that's been the backbone of what we've wanted to do in the first place.

And it's thankfully been good. The benefit of us having done it now is we've shown the absurdity of the copyright laws as to, you know, when somebody thinks that all my muse gave me this melody and this is really came out of nowhere and this is out of my heart, blood, sweat, and tears that this melody came. And we've kind of said, well, dude, it's just math, right? I mean, you picked out of all the mathematical permutations. You picked out one of those permutations. And yes, it's beautiful, but it's not unique to you.

And if somebody else picks out the same beautiful melody, it's not because they copied you, dude. It's just because there are only so many notes and we're running out of them. As people are uploading into Spotify at the rate of, you know, 100,000 per day, of course we're going to overlap because we have to, right? So that's really the goal of this is that, one, we didn't want nefarious bad people to do it, and two, that we are running out of spaces and we wanted to shed light through the largest platform that we could, which is TED Talks, to be able to say, hey, maybe we should stop seeing each other over these things. Think about the last time you heard a breach story on this show.

It always starts the same way. Someone somewhere saw something too late, an alert buried, a signal missed, an SOC that just couldn't keep up. Arctic Wolf set out to solve that problem by rebuilding security operations from the ground up for a world where attackers are already using AI. They created the Aurora superintelligence platform, a fully agenic system powered by the swarm of experts. Instead of single-purpose bots or lucky-guess LLMs, this swarm is full of deterministic agents that handle whole entire workflows. Humans stay in the loop and on the loop to validate the critical decisions and keep everything trustworthy. And all of this is just off running on their secure operations graph. A constantly updating intelligence engine fueled by more than 9 trillion telemetry events every week

and over a decade of real-world incident response. The system reasons on real signals and real context, not synthetic training data. And the result is the new Aurora Agent SOC. It's the first SOC that is agent led by design. You get agents that coordinate, agents that investigate, agents that respond at machines. speed and hundreds more that automate the repetitive work that normally buries human analysts. Arctic Wolf didn't try and bolt AI onto an old model. They rebuilt the model entirely. What makes it even more effective is how it works with Arctic Wolf's concierge experience.

The team brings customer-specific context directly into the platform so every AI-driven decision reflects your environment instead of generic assumptions. The automation frees your concierge security team to focus on higher value strategy and proactive risk reductions. while the agents handle the grind. If you want to see what trustworthy, production-ready AI and security operations actually looks like, go to arcticwolf.com slash hacked. Never feel like cyber threats are evolving faster than anyone can keep up? Last year, 2025 was nothing short of a record-breaking year for major breaches,

from sophisticated ransomware operators to AI-enabled attacks that turned defenses on their head. Organizations around the world saw headlines they never expected, than cybersecurity teams were tested like never before. But here's the thing. These incidents aren't just news headlines. They're learning opportunities. And that's why Arctic Wolf is hosting a live webinar on February 5th, diving the most impactful breaches of 2025.

Their field CTO and security leaders are going to unpack not just what happened, but why these attacks succeeded. And most importantly, what businesses can do to fortify their defenses for it's too late. You're going to walk away with real insights into how threat actors are evolving, how defenders are responding, and what strategies can help you stay ahead of the next big breach. It's not fearmongering. It's practical, actionable, intelligence from experts in the trenches.

Register now at arcticwolf.com slash hacked. I want to talk more about the melodies as math thing because I find it so interesting. But do you think there is a, is there a type of ownership over a piece of music that makes sense to you? Obviously, melody to you, and I think I would agree, a melody is math, but a full arrangement, melody plus words. Where is that line between sort of common knowledge and I created something? The muse did strike me and I created this. I'm so glad you asked that. Yeah, a song is 100% copyrightable. Yeah, we should totally

have copyright over songs. Because if somebody says, you know, it takes any Beatles song and then just puts separate words on it, right? Of course the Beatles can sue, right? That is a song that we are going to go through. Really, that is not what we're addressing. We're addressing just that one component of the song. Melody is a component. Harmony is a component. Rhythm is a melody. If there are five timbre, right, there are all those elements. All of those together constitute a song and are therefore copyrightable.

But separate any of those elements individually. And if you try to sue over any one of those elements, that's where I say, hey, dude, drumbeats, not copyrightable. Chords, not copyrightable. And now maybe melodies, also in isolation, maybe not copyrightable. Interesting. What else? It's weird to say. what else is math that we don't think of as being math?

Things that feel like really elemental parts of creation that are a lot more like trigonometry. Yeah, so a lot of it deals with combinatorial mathematics. So what that means is the mathematics on my project, I told you the doremithosol litigio, that's eight up, and then repeated 10 times or 12 times. The mathematics on that is 8 to the 12th power. Okay, so that's that 8 to the 12th power is 68 billion.

So there are only 68 billion, you know, eight octaves with 12 note repeats. To make a simpler one, if you do one, two, three, two, one, right? One, two, three, two, one, that is three notes up and one, two, three, four, five to the fifth power. So what is three to the fifth power? We'll ask from the Google. three to the fifth power is 243. So there are only 243, three notes up, one, two, three, and then one, two, three, four, five,

like that. There are only 243 combinations of those. So how fast are we going to run out of those three up, five across? Really effing fast, right, in a fraction of a second as far as Spotify. So really, when you ask, is there anything that is like music, that is comparatorially small, Contrast what I've just talked about with the number of words in the English language. So for example, the words, number of octave is eight, but words in the English language is 117,000.

And then if you have how many five word phrases in the English language, the mathematics on that is 117,000 to the fifth power. So if we do that, 117,000 to the fifth power, that is, there are 25 zeros behind that number. So there are probably there were fewer sands, grains of sand on the earth than there are that many combinations of that. So really mathematics, music is a very, I think, unique in that there are only a small number of ways that you can express the Western 12-tone diatonic or chromatic scales. And that's why it's relatively unique in art. Have you heard about, just because you brought up words, have you heard about a project, it's the Library of Babel? of Babel. Yeah. So cool. So for anyone listening that doesn't know about this, this is a website inspired by the short story, the library of Babel, about an infinite library. And the website

is a simulation of this infinite library. It's divided into, like the metaphor on the site is it's divided into hexagonal chambers, each with four walls, bookcases, five shelves per wall, and 32 books on each shelf. And you can navigate them. And you can go into the pages of each book. and in one of those rooms, you can find pretty much any piece of text. It's not generating the text. The text is always waiting on the same page of the same book in the same spot. I checked this before our conversation. In page 195, volume 31, shelf four, wall one, it reads, on September 7th,

2023, Jordan Blumen interviews Damien Rale. That's just waiting there. That's been sitting there since 2015 when the project was created. The creator of that site authored this. infinite text. You and Noah are kind of authors of this, not infinite text, a few billion strong text. And just as a layperson, it feels like the creators of chat GPT, the creators of mid-journey, the creators of those things sort of created these almost infinite texts as well. You spoke to this, but what responsibility to authors of such

texts machines call it what you want? What responsibility do they have to the rest of us? That's a really good question that I've given many interviews. No one's ever asked that question. So kudos to you on asking a very original question. There is creativity after all in creativity brands. So I would say to the first of the library Babel, yes, there is an aspect of that, that he is doing that algorithmically. That is that the text that you put in there on September, on this date, September 7th,

Damon spoke to Jordan. That technically didn't exist before. you did it, but he said this is where it would have lived. This is where it would have lived had it been in the infinite. Because you can imagine that infinite text would take up terabytes and terabytes and petabytes, and it would just be too massive. So what he's done is created an algorithm to be able to be able to say, this is how all of them exist. And I can mathematically prove that all of them here. I can mathematically prove that through the algorithm. And then each of them has a location. So essentially by you putting in the input, it's saying where in that location they are. would have been.

Right. And so somebody said, hey, why didn't you and Noah do that on our side? But under copyright law, this kind of theoretical location isn't good enough. Within copyright law, it has to be fixed in a tangible media. That is, it has to be fixed to a hard drive. So we actually had to write out the particular notes onto disk to be able to actually reflect them. So anyway, so that's thing number one.

Library Babel is cool, but it's kind of a parlor trick a little bit because it doesn't actually exist. It's just, it is showing you the actual location had he created the terabytes worth of data. Thing number two is what is the responsibility of, you know, large language models like chat GPT and Anthropic and Google and others that are building this to ingest the entire internet and then to be able to say to humanity, do we have an obligation to be able to, I guess is the question to use it for good and not evil? Is that the question? I think that's probably almost always my question whenever I'm thinking about these things.

Yeah, I think that's the question. Yeah, I think that, of course, hammers could be used to build houses and they can be used to hit you over the head and kill you with the hammer, right? So they could be used for good and for ill. And so I think that to the extent that the hammer makers can keep the hammer from hitting people over the head, I think they should. But I think that there's a larger question as to society is a large language model more like a Ford Pinto. That was if you got rear-ended, it would burst into flames. Or is it more like a hammer where you just like, hey, that's the risk of having hammers, right? is that people can sometimes hit them each other over the head with them.

So I think the jury is still out as to large language models and the extent to which we as a society have an appetite to regulate them like Ford Pintoes and to be able to say here, you know, make sure you put guard rails on them to be able to ensure that they're not being used for to sway elections or to create misinformation campaigns and that kind of thing. And I think that I think that danger is strong enough where I tend to lean to more of the Ford Pinto side that maybe we should regulate them. But do I trust our legislators to know anything that you and I have been talking about over the last 24 minutes?

And the answer is no. So I think that the odds of them being able to regulate anyone in the world to be able to regulate the large language models with effectiveness is very slim. Just using the last couple minutes I have here with you, Damien. You have such an interesting intersection of law, creativity, and tech in your career. And I imagine you bump into a lot of misconceptions. I imagine you bump into a lot of people that maybe have a professional responsibility to know more about how these things intersect, but don't. People are busy. So if you could leave people with almost like one misconception you'd want to bust about that intersection of human creativity, technology, and the laws that govern these things, what would you want to tell people?

What is the thing that you bump into most that you would like to correct? I think that I'm a musician, you're a musician, many of us musicians, think of our creations as our babies, right? And if someone, we think about we as musicians have been kicked for the last 30 years, where we're making less money than we've ever made. We can't make any money from recording, right? We can't make it, we make less money from touring, right? So a lot of us think of, well, like maybe if I'm going to make money, maybe it's through copyright. So if there has been an objection to my thing, they're saying, hey man, Damien, you're taking away even my copyright, right? I can't even have my copyright anymore. How am I going to make any money?

So that misconception, to those people, I would say, the thing that you're building is more than the sequence of notes that is reflected in my database of 471 billion melodies. All it is is a sequence of notes. You are putting the ache of your voice. You're putting the whale of your guitar. None of those things is reflected in my 471 billion MIDI notes that are on a hard drive somewhere. And that is 100% of the reason that your audience likes listening to you is because of the ache of your voice and the whale of your guitar. So the thing that I'm taking away is not any of your well-being. What I'm taking away is the target off of your back, that if you actually make it big, that's somebody who uploads one of the 300,000 melodies that goes up to Spotify every day,

that maybe some of them hit it big enough and maybe that melody is the same as yours, and maybe you get sued over a song you've never heard before. What I'm taking away is the for that person to sue because hopefully a judge will say, hey, there are only so many notes, and we're running out. Okay, actual last question. On that hard drive, all those melodies, do you have a favorite? Have you ever heard one that just made your heart sing? That's really funny. To be honest, almost all the melodies suck. Because I mean, who's going to listen to? I love that one. Right. So good. So really the question is, you know, the 471 billion, the number of listenable, actually pleasing, aesthetically pleasing melodies,

are very small, right? And so then, so people say, 471 billion, that's a massive number. When are we going to run out? I'm like, well, all all of them sucks. So the 95% that suck, there's only 5% of those that are actually pretty good. How fast are we going to run out of those pretty quickly? So to answer your question, I don't really have a favorite melody. I guess I could make up a Beatles song or something, but I would say that there's a lot of shitty melodies I hate. If it's such a small subsection of them that are actually good, could you find it, could you train something to parse through the massive database to find the ones that share the positive qualities or qualities we think of as being good melodies and kind of suss them out?

Be like, hey, just so you know, there's this banger right over here. No one's even found it yet. Could you train something to go digging? I'm funny you asked that because there's a Spotify has a patent that has not yet been granted, but it has been applied for, where if you input a melody, it will output all of the songs that use that melody. And so I've actually reached out, one of my friends is the former cheap economist of Spotify. And I said, hey, could you put me in touch with the inventors within Spotify of this thing? Because what I want to do is anyone who's seen my TED Talk knows that there's a grid of every potential melody in there. And so think of that as all the white spaces.

What that Spotify patent could do is fill in those white spaces. with melodies that have already been taken. And then tell me these are the blank spaces that haven't been taken. Here are the red spaces that have been taken. And more importantly, here are the gray spaces that Bach and Beethoven and Mozart and everything that's already public domain have also taken. So I said, wouldn't it be great to take my 471 billion dataset, fill in all the black, all the red, all the white, and all the gray spots,

and be able to say, this is what's left. and of what's left, hear the bangers that are doing. But more importantly, here are all the copyrighted songs today that are ripping off Bach and are ripping off Mozart and ripping off Beethoven. There are only so many notes. And of course, all of the bangers out there today have been ripping off all of these people that have been dead for hundreds of years.

So how special is your melody if Bach and Beethoven and Mozart did it hundreds of years ago? I really appreciate you taking the time to shout with me, Damien. Jordan, it's been a pleasure. Really, thank you for having me out.