In this episode, Ezra Sandzer-Bell shares his journey from classical and math rock musician to software developer, educator, and founder of AudioCipher — a tool that turns words into melodies and chord progressions. Ezra discusses how cryptographic techniques and secret codes have historically appeared in music, from Bach to 90s post-rock, and how these ideas inspired his books.

We explore the philosophical and technical ideas behind AudioCipher, how it reclaims agency in an AI-dominated landscape, and how it fits into the broader question of creativity, automation, and authorship in music today.

Ezra also talks about his role as a writer and marketer in the AI music space, the dangers of generative audio flooding public spaces, and why honesty and critical product reviews are essential right now.

So hi and welcome Esra Sanserbel.

2

I hope that's a decent enough pronunciation.

3

I don't actually know how you say it.

4

You did a great job.

5

Thanks.

6

I'll do a short introduction of you.

7

So Ezra Sanserbell is a musician, educator and founder of AudioCypher, a plugin that turns

words into MIDI melodies.

8

And you started out writing books about cryptographic music theory and went on to teach

producers music theory through interactive tools like hook theory.

9

And today you write widely read articles about AI and music.

10

consult startups and develop new tools that explore the creative potential of generative

tech.

11

Does that sound about right?

12

That sounds right.

13

Yeah, you did your homework.

14

Great.

15

And I was thinking it would be nice to hear a bit about your journey as a start.

16

how did you start doing music and how did you end up going more into the tech side of

things?

17

Hmm.

18

Yeah,

19

Well, my family, everybody on both sides is a musician.

20

So my dad plays piano, my mom plays flute.

21

I grew up trained on cello, but I grew up listening to my dad play piano.

22

And even my mother's father, my grandfather was married to a woman who was a classical

pianist who would play Debussy and Chopin.

23

And so I grew up with kind of more of a classical and jazz influence.

24

Like my parents weren't that into rock and roll.

25

And so when I was like a teenager, I started getting more into rock music.

26

I went from playing cello in the school symphony orchestra to

27

playing bass guitar and then eventually guitar.

28

When I was in high school, I played in a math rock band.

29

Math rock is kind of like, you could say like a polyrhythmic or odd time signature

30

Maybe you could say like late 90s.

31

And so that was kind of like where I got my start as a musician.

32

I put out a couple of records with that band and we toured and created a name for

ourselves.

33

is still selling records and still drive some monthly Spotify revenue, which is cool.

34

And so yeah, I kind of had like my first aha moment as a...

35

just a creator there because rather than just like playing other people's music and in

orchestra, I was getting together with a couple of guys a few times a week and practicing

36

a lot at home and creating something from scratch.

37

And so it's like the entrepreneurial spirit of the DIY punk scene kind of, you know, mixed

with more of my like classical roots.

38

Eventually in my early 20s, as that band started to fizzle out, I went back to college and

did four semesters of music theory and got like my more fundamentals under my belt.

39

And as I was going through that process, I kind of hit this wall where a lot of my innate

creativity was starting to get stifled by sort of...

40

I guess, ideas about music.

41

Like the way that they take you through university programs that you're interviewing me

from a university, so I'm sure you have some exposure to this.

42

It's very much oriented towards like traditional Western canon.

43

So Bach and you have to do four part chorales and voice leading and all these principles,

which are important, which I understood more intuitively.

44

through my hands and through my ears and later on through my college education was more

being sort of imprinted on me through these classical forms, which is helpful.

45

I actually think that like learning structure of harmony and melody is great.

46

And it like did something to me creatively that like shut me down, which was really

interesting.

47

So then there was this whole unlearning and undoing process.

48

And so that was like around 2007 I started.

49

At around that same time YouTube kind of came into popularity.

50

And so I started like finding, you know, I guess you could say like unusual YouTube

content around like a variety of different topics related to, you know, Western

51

philosophical traditions like alchemy.

52

Alchemy is kind of this buzzword that we hear thrown around a lot, especially in music

circles.

53

But, you know, there's like a really rich hermetic tradition in Europe that, you know, was

intimately connected to the origins of Western music.

54

so.

55

You know, I just became really interested in this part of history that usually isn't

taught in university programs, at least not in the early stages of the bachelor program.

56

So I did a lot of independent research.

57

There's a professor out of New York named Jocelyn Godwin, and he's like a he speaks many

different languages and has translated all of these important philosophical texts about,

58

you know, astronomers and like their belief about like how music and astronomy were

connected.

59

And so I started kind of like formulating my own ideas around how those were connected to

my own personal experience as a songwriter and as a musician.

60

And so as that stuff is like cooking and I'm like taking notes, I come across this one

tiny idea, which was that Johann Sebastian Bach in the Well-Tempered Clavier, which is

61

kind of like, you know.

62

the Bible of the Western canon.

63

He had used this thing called the Bach motif, which was a musical signature.

64

So basically, maybe you know this actually, because you're in Norway, but in Germany, they

actually have A, B, C, D, E, F, G, H.

65

And H is what we call B and B is B flat.

66

Okay, yeah.

67

So B, A, C, H actually spelled out what we would call like B, A, C, B, if I'm not

mistaken.

68

And so,

69

So that four note chromatic motif is actually, you could think of it as like a watermark

on a bill.

70

Like it's not visible until you hold it up to the light and you can kind of see like

invisible ink, you know?

71

So basically the musicologist who would study the wild tempered clavier noticed this motif

woven into.

72

pieces within those published works and there were other clues that he had left throughout

his life that indicated that that was an important motif.

73

And so when I read about that and then I discovered actually there had been, you know,

centuries after that, this whole tradition of musical cryptograms used by composers where

74

they would embed their name into melodies, I just thought, wow, like what a weird thing,

you know?

75

what, it's like I've never heard of that.

76

And then, know, other sort of like signals started popping up for me.

77

Like there's a post-rock band from the late 90s called Hazel.

78

They were on Sub Pop, I think.

79

I could be wrong about that.

80

But they were like a pretty...

81

You think about Nirvana.

82

It's kind of like a Nirvana era style band, you could say.

83

And they had this song called Comet where the chorus was, and everywhere comets flare.

84

And in the music video...

85

there people were sign language AECF and there was literally like a ad lib in the

background.

86

The guy goes AECF, which is the first letter of each word in that chorus and everywhere

called its flair.

87

And I could hear these like four power chords under it and I go, there's no way they're

playing AECF power chords.

88

Like that would just be too perfect.

89

And I listened and they were and I realized, musical cryptograms are even being used by

rock artists in the nineties.

90

So then I started thinking like, is this like a whole hidden subtext to music that I've

totally overlooked?

91

I thought about, you know, John Williams and the Close Encounters of the Third Kind with,

you know, Spielberg's famous movie about the UFO communicating with humans and that five

92

note melody.

93

I just was so curious, you know, it really, it struck something in my imagination that

connected into my love of music theory and my desire to like combine.

94

the early stages of imagination and creativity from being in a band and being like kind of

a solo singer songwriter with this like new part of me that had been installed for my

95

university education, which was like, you know, just conventional four part chorale

writing and music theory.

96

And so that's really, you know, it got me into that.

97

And I found out about ciphers.

98

I found out that there were spies and people involved in espionage, like as early as the

17th and 18th centuries who were using musical sequences

99

to send secret messages to one another.

100

And so essentially like what would appear to just be sheet music actually have like a

decoder.

101

And there's many different instances of this over time that are like recorded in

encyclopedias.

102

And so I realized, it's not just for composing music, like, know, secret messages, which,

you know, to evade detection in order to like communicate something like that felt also

103

like it had this sort of like.

104

interest for me was like, you know, messages encoded in melodies for music, but also

messages encoded in melodies for just communications.

105

So all this stuff then it started tying into my understanding of encryption, because

really encryption is the whole basis of internet protocol.

106

mean, and you know, obviously, like, you think about crypto is, you know, the basis of

that is something called cypherpunk.

107

Cypherpunk was like a

108

a movement that predates Bitcoin where it was essentially people saying, yeah, you know,

we need a trust, a trustless system that we can build a new currency off of because fiat

109

currency and the global economy is like tied into all these power structures, yada, yada.

110

So I was just like, man, it's like everything around like codes and, you know, music and

all this stuff is so fascinating to me.

111

And then I started noticing in movies, like movies and video games.

112

Like I remembered back to all these things that I had loved, like

113

The Ocarina of Time is a famous Zelda game and like, you you play this melody on a flute

and it has some magical effect on reality.

114

It changes night to day or it causes the wind to move in a different direction.

115

I think of Batman Returns playing a three note melody on his piano in the back, the

Batcave opens up, you know, he walks through his library door down into the Batcave.

116

I mean, just dozens of examples of this from, from famous movies.

117

And so

118

What emerged for me was like this greater Gestalt, this this holistic view of music as a

language, not just in the emotional sense and in the interpersonal sort of human to human

119

connection sense, but literally as like a medium through which like a second layer of

meaning, like an entire other world of like language could be encoded into melodies and

120

chord progressions.

121

So in 2014, I published my first book, which was called Astro Music, which was like

really, I mean, kind of a punk DIY approach to writing about this stuff.

122

had no publisher.

123

Actually, that's not true.

124

It was published by a company called Sync Book Press, which was a legitimate publisher.

125

at the time when I wrote it, I didn't get an advance.

126

So I was just working a full-time job at a bakery, typing out every day in my free time.

127

all my ideas and trying to synthesize this stuff.

128

Eventually I published a second book with them called Audio Mancy.

129

And so the whole premise of these books was like music has this fascinating sort of

subcurrent of linguistic meaning that like nobody is even talking about or even aware of.

130

And so after publishing those books, you know, I did like, I drove up and down the West

coast of the U S and was going to bookstore to bookstore selling my books.

131

And so I had this sort of like entrepreneurial moment.

132

where I had to kind of like, you know, get people to buy in.

133

And eventually I was speaking at, you know, book conferences in Portland and up in

Seattle.

134

The biggest book conference I had was for about seven or 800 people in Seattle.

135

And I did like a book signing after, and then I was given a book deal offer by Wiser,

which is like a big sort of like new age esoteric book publishing company.

136

And I just thought, you know, this feels like

137

If I go down the route of like being an author and writing books, I'm going to be trapped

in like this print medium.

138

And I could just sign a, see the writing on the wall that like, wasn't a career where I'd

really be able to like make the kind of living that I wanted.

139

And, yeah.

140

And I didn't have like a real career.

141

mean, I was still just working these like close to minimum wage, like restaurant jobs, you

know, and, playing in like bands and writing music on the side, but like,

142

I just didn't feel like I'd figured it out yet.

143

Around that time in 2016, I met some digital nomads in Thailand.

144

Actually, I took the little savings I had left and flew out to an island in the south of

Thailand.

145

I was just living in this literally like a bamboo bungalow for like $3 a night.

146

But there was an internet cafe thing next door.

147

And so I'd walk there every day and I met these guys who had these Amazon drop shipping

companies who were making like $30,000 a month.

148

And I just couldn't believe that you could make money off the internet like that.

149

It just blew my mind, you know?

150

And I met this one guy who was, he was from London.

151

He was like a UK based DJ who had bought this mailing list through like a friend of his

who also did music stuff.

152

And he was just creating these like new age, like meditations.

153

but there would be like a one hour meditation thing and he'd sell it for like 89.99 or

something.

154

And I was like, how are you like, who's buying this?

155

And he goes, no, he goes, I put all this love and intention into it.

156

And I have a really, you know, I spent all this time with my readers.

157

And so I was like, wow, like that's a, that's a crazy thing.

158

And he goes, you know, we should try to sell your books as, as PDFs.

159

And we did that and we actually made some money.

160

It was my first aha moment of like,

161

Like I can sell my knowledge through my written work through like mailing lists.

162

But I didn't have a mailing list and I couldn't buy his.

163

So I kind of had this like brief moment of awareness followed by like seven years of not

doing that.

164

Basically what happens after Thailand, I went back to the States.

165

The little money I had left I invested in going to code school.

166

I took a six month bootcamp for programming.

167

I learned to code.

168

And then I took about a year basically applying for jobs.

169

And I got a job at a software company in Portland where I worked for five years.

170

And it was a small startup doing appointment scheduling.

171

And I wore all the hats.

172

I did everything you can do at a startup.

173

And we went from about 300K in annual revenue to about 1.2 million.

174

So we pretty much forexed the revenue with just me and the founder and one other guy.

175

And so that was, again, like another huge

176

learning moment for me of like, my communication skills and my writing can like earn me a

serious salary at a software company.

177

It's not music, but it's that.

178

So in 2020, then I got COVID hit and my partner at the time was a designer.

179

And we just had this thought one day we're like, man, we're so bored and like everything

shut down.

180

Let's just, let's start a company.

181

Why not?

182

And so we started audio cipher and it was essentially my attempt to take all the stuff

that I knew about and just make it like a tool, like a discrete tool that anybody could

183

use that would turn words into melodies.

184

So V1 was words to melody, V2 was words to melodies and chords with some rhythm, V3 was

like, or no, I'm sorry, V2 was melody with more rhythm and more scales.

185

V3 had chords, V4, which is the latest version that came out in September, 2024.

186

has chords, melodies, and all sorts of randomization and inversions and voice leading and

note joining and really the most advanced algorithm that we could develop.

187

But it also has something called the MIDI Vault, which is a MIDI file management tool.

188

So essentially you can save any cryptogram that you create, any chord progression or

melody you create in a kind of card with metadata, which has like BPM, key signature,

189

instrument type, mood, genre.

190

a rating, a card name, notes, et cetera.

191

So it's like everything you might want to have.

192

And then, you you can filter through all of that data in the MIDI vault.

193

And the main inspiration there was that you have these MIDI files and audio files that can

be stored in one container, which means that if you have, you know, a MIDI idea and you

194

apply sound design and you bounce those audio tracks and import them into the vault.

195

then you can have the samples in maybe a few different styles, and you can have the

original blueprint in the MIDI.

196

So next time when you open up a DAW project, you're like, hmm, I'm in this mood today, or

I wanna write for this genre, you go through your collection, and you can actually just

197

pinpoint what you want, and you've got the MIDI files if you wanna add different sound

design, or you've got the samples if you wanna preview what it should sound like.

198

Also, you can export your collections, and you can import them.

199

So there's a trading.

200

component here, which is a bit like trading cards with your friends.

201

So let's say you and I were working on a music project.

202

I'm like, Steiner, I've got like these 12, you know, new ideas.

203

Let's make a song together.

204

I just send you the audio cipher zip.

205

You can import that, that AUCI file, and it just loads automatically with all the metadata

and the audio files exactly as it was on my computer.

206

And so it's a collaboration tool.

207

And so I feel like that kind of gets you up to speed, but that's where I am today.

208

You know, what you're seeing, I think the way we met on LinkedIn is that I'm also acting

in the capacity as a music marketer.

209

And that's because, you know, over the last two years, as I was scaling up the company, my

primary growth engine was organic.

210

In other words, my blog articles were driving 50 to 60,000 unique site visitors a month.

211

Just again, like through my writing and some percentage of those people then go check out

our homepage and some percentage of them buy.

212

And so it was enough to kind of, you know, keep our company running and

213

when I took that to other companies and showed them what I had done, they were like, can

you recreate that for us essentially, which I did.

214

So I've worked with a number of AI music companies, NeuroFrames, SunoAI is a big name a

lot of people will recognize.

215

I've worked with RIPX, I've worked with Hook Theory is one you mentioned.

216

I mean, probably a couple dozen at this point.

217

I still have a few clients that I'm working with right now.

218

And maybe this is, you know, I'd love to open the space for you if you have any thoughts

or questions on this stuff.

219

But the thing that maybe connects in with you and I know you're working in education is

one of my clients, they're called Future Proof Music School.

220

They've just recently launched, but they actually have a AI chatbot that can listen.

221

So it's actually able to analyze sound down to the mix and make like expert

recommendations on how to improve your mix.

222

And so...

223

you know, something I left out was I worked for two Ableton certified schools, one in San

Francisco called Pyramind and a second one called Champion Sound, where I worked combined

224

about like probably two and a half, three years.

225

And I would do these classes like either one-on-one mentorship or I would do groups of

like six people at a time.

226

So I really have an intimate relationship with like teaching.

227

Both my parents were university professors, by the way, at different phases of their

career.

228

And so, yeah, I feel like my mission in life is

229

you know, to create something that's inspiring and beautiful, to, you know, build product,

to make music, but also to teach and to make sure to remove the barriers to discovery that

230

I feel like some of the industry gatekeepers can kind of throw up.

231

And so, you know, what I'm doing with my clients essentially is helping them to tell their

story, you know, find their position in the market, make sure that they're

232

creating content that people will understand and that's helpful and makes it easy to

understand how to use the product.

233

so, yeah, that's where I'm at currently today.

234

Alright, thanks so much.

235

No wonder you're a good writer, because not many people I know can keep a flow of a story

going for so long and so concise and clear as you just did.

236

And it was really fascinating to just sit here and listen.

237

So basically, you've been through this whole process of connecting different interests.

238

And it was really fascinating when you started talking about cryptography and Bach and you

unloaded a lot of information that was completely new to me right there, which was

239

fascinating.

240

So as I understood it, you were saying that Bach interweaved his name in the music through

the...

241

as simple as just the notes B, A, C and H.

242

So that's B flat, A and C and then B in English basically.

243

So that's kind of a way that a musician could put his signature in the music as a visual

artist would sign his picture.

244

He could just sign it with his initials.

245

That's fascinating.

246

And that kind of sparked another.

247

well, one thing I want to say for you is that, you know, obviously you run out of H.

248

What do you do?

249

So what the French and European system is, I think that is called the English French and

English cipher, which was.

250

it's more than 100 years old now, basically when you get to G, the H starts back at A

again.

251

So A, B, D, E, F, G, H, J, K, L, N, O, Q, R, S, T.

252

So you can kind of imagine how that goes.

253

And so you end up with just a simple seven column matrix.

254

And that was the foundation, that is the foundation of audio cipher.

255

That sort of historic thing.

256

so musical cryptograms are very different from what most people think of with text to

music.

257

We think about

258

description parody and well I typed in a lo-fi hip-hop beat with xyz instruments and

that's not what I'm getting back well yeah that's not the intended use of it you know so

259

it's it's one of the things that's interesting about audio cipher is it really disrupts

the the trance that we're all in that like text turns into the thing by way of ai taking

260

control and doing everything for you the vision with audio cipher

261

and which I've done like dozens of demo of this on our site that anybody could check out,

is to take a word like frog and then apply sound design that sounds froggy and have some

262

croaking sound effects and create the atmosphere and the mood of a toad.

263

And what does frog synth sound like?

264

Right?

265

So it's actually audio cypher is not about handing over creative agency to AI.

266

It's about

267

harnessing this thing that's implicitly interesting to us as creators, which is turning

text into music, but retaining all of the creative control, except for one little

268

constraint, which is that the letters of each word are encoded into that melody or chord

progression, which means you never would have arrived at that specific sequence had you

269

not used that word.

270

And so in a way, the word is forever connected to that piece of music, which is kind of

fun.

271

Definitely.

272

And we'll talk more about AI and music after a while.

273

I find it also interesting that a common theme in the AI music space is that a possible

negative consequence of using AI to generate music is that you kind of lose the struggle.

274

And the struggle to make art is sometimes...

275

typically as the art itself.

276

So the struggle is the art.

277

And through audio cipher, you still get that struggle.

278

It's more like a confinement that lets you be creative inside the confinements of that

cryptogram or that overarching concept, which a frog might be.

279

I teach composition at the university and one year I gave my students a really weird

assignment.

280

where I just told them to meet up at nine at this particular spot.

281

And I didn't really say anything other than that.

282

And then they met up.

283

I arrived and I took a water bottle and I just poured the contents of the water bottle on

the ground.

284

And then I said, okay, now we have what you need.

285

Make some music.

286

So that's a bit similar to the frog experiment, I guess.

287

So it makes me wonder if I could do some sort of composition or arrangement.

288

assignments including audio cipher.

289

think that would be fun.

290

Before we move on could you tell us some examples how people use audio cipher?

291

Sure.

292

Yeah.

293

Well, we talked about the frog example.

294

mean, there, you know, there's a few use cases that I think appeal to people.

295

So, you know, one of them you've seen, maybe, for example, Suno is like a really popular

AI music tool.

296

And one of the things that they really led with was like create music for your grandma and

grandpa or for your girlfriend or boyfriend or for your pet.

297

You know, and so I think the idea of devotional music sounds like really heavy and kind of

like spiritual or religious.

298

But I think devotional music really is about like I made this music

299

for you.

300

And like there's something really powerful about that.

301

So I think one of the coolest ways to use audio cipher is to make music for something you

like or something you love.

302

And that could be anything.

303

It could be a person or place or thing.

304

It could be a slice of pizza.

305

I mean, it doesn't matter what it is, you know.

306

But I think that to me, that's my favorite use case for it.

307

And then also, you know, there's the midi file management piece of it, which we, you know,

aren't focusing on as much, but that's a huge thing, which is sample management.

308

mean, people have all these DAW project files that are just like some random sketch that

they just saved and they never go back to.

309

And there's hundreds of memes on Instagram.

310

I mean, I started liking these memes that joke about this issue and I realized it's like,

it's just an endless joke because we all have these like poorly named files.

311

If you want to search through it, it's like the file name is like

312

whatever random thing dash the BPM dash the key dash the instrument dash the genre it's

313

our card system was intended to solve that problem, but also connecting MIDI to audio.

314

was like the really the most innovative thing we did.

315

There's other sample managers out there.

316

Some of them have more traditional kind of views, you know, but I think the way that we

went about it by combining that tool with the text to MIDI generator, it just wound up

317

being kind of a unique, you know, combination and the bridge between the two of them is a

save button.

318

So basically you can save any

319

text-emitting creation you generate and put that right into a card.

320

Nice.

321

And I'll put a link to to all decipher in the article about this episode so people can

check it out.

322

So I found you through through LinkedIn and on LinkedIn you you do a lot of takes on on AI

and music, both on the skeptical side and also on the optimistic side.

323

So I thought we might explore both sides of the argument.

324

Hmm.

325

So let's start on the negative side.

326

What examples of generative AI in music generally provoke or frustrate you these days?

327

I find myself in this unfortunate position of being the grumpy guy in middle of town

square ringing the bell going, this doesn't work, this isn't good enough.

328

And I don't take pleasure from that actually because I'm generally, I like having good

connections and relationships with the people in my industry.

329

The thing that frustrates me the most is hype that is disconnected from product.

330

And so at the end of the day, I've committed myself for better or for worse, probably at

my own detriment, but I think for the benefit of the greater software and professional

331

community, hopefully, to writing honest reviews about my actual experience with product.

332

So, you know, and beyond just my frustration with hype, would say, you know,

333

people are rushing to get stuff to market as fast as possible, you move fast and break

things as we say in tech.

334

And that's resulted in some crazy bugs.

335

I the worst one, I'm not gonna throw them under the bus and name the company, but there

was one company I was quite excited to use their AI plugin.

336

And there was a critical bug that actually ran a system override on my speakers and blew

out my right headphone.

337

And I had tinnitus for like three days and I DM the founder and was like, hey,

338

What the heck and he was like, sorry about that.

339

Like we we just urgently fixed the bug Would you like to try it again?

340

I was like no you guys need to QA your stuff, you know So yeah, I think like a lack of

quality assurance and like rigorous testing.

341

It's just like it's actually the case in a lot of you know smaller product, I mean

342

The unfortunate reality, and I say this as an insider, as a software founder myself, who's

also worked, like I said, with dozens of other companies and, know.

343

and CEOs there is that there's just not a lot of money in the space.

344

know, mean, companies are working on lean budgets compared to other industries, know, like

enterprise B2B industries where there's just millions of dollars to work with.

345

I mean, there are exceptions, know, Suno, UDO, Refusion companies to get, you know, either

a few million dollars or in Suno's case, $125 million of VC to play with.

346

But like the majority of us, I mean, I can speak as audio cipher and we're totally

bootstrapped company.

347

We're not making that much money per month.

348

And so there's not that much budget.

349

I'm still, you know, I'm still have not turned a profit for myself after five years.

350

And we've done over six figures in revenue, but the cost of a good designer and a good

developer and paying for marketing and paying for everything, you know, I it all comes out

351

of pocket.

352

So I have a certain compassion for these founders who are shipping product as fast as

possible.

353

It's not that I intend to like put, you know, put them in a bad situation by calling out.

354

when the product's not functioning as it's hyped up to.

355

But at the same time, if there's no accountability, then there's no real transparency in

what happens.

356

We've seen this not just in music, by the way, it happens in all the different AI news

stories is like one person claims that something is the case.

357

And then people just repeat that ad infinitum and nobody takes the 30 minutes of due

diligence to just check and see if that's even true.

358

And so that's the thing that frustrates me the most, you this, this, this stuff that most

people complain about with AI music and AI audio is like the disruption, the unfairness to

359

artists, like all this stuff.

360

It's like, yeah, I'm with you on all that.

361

But like, to me, it's, mean, I just live and breathe this topic.

362

like,

363

I've come to a certain point where I'm like, accept that that's part of this

transformative experience we're having as a global civilization.

364

AI is extremely disruptive.

365

And I don't see that many paths for us.

366

One of us is to bare our heads in the sand.

367

Another one is to scream from the rooftops.

368

Another one is to just blindly hype and promote.

369

mean, there's these sort of like lanes that we're all kind of jettisoned into.

370

And I just feel like, well, the best thing, like the healthiest thing that we can do is to

use the product and understand it and criticize it.

371

Not just criticize it on an ethical and moral high ground, but to criticize where it's not

functioning properly.

372

If this stuff is supposed to be an assistant for humans and not just take over our jobs.

373

We have to become intimately familiar with it.

374

Know thine enemy, so to speak.

375

You know what I mean?

376

And so anyway, yeah.

377

I totally agree and I find that you do a great service in doing honest reviews because

that's kind of lacking in the space, I think.

378

I try to follow along with what's new each week or each month in regards to plugins or new

software.

379

And I would say like 99 % of what's being hyped doesn't function as advertised or it's

380

simply low quality sound or low quality interface or yeah so it's really helpful to have

someone doing honest reviews and I guess the plugins or companies themselves also would be

381

appreciative of such a reviewer when they finally have a product that is good because then

you have a reliable source of marketing I guess you could say.

382

Totally.

383

I've also read a post from you a few weeks or months ago, I think, where you noticed

generative music being played in cafes and restaurants.

384

And that's something that's a bit provocative to me, at least as a musician.

385

How did you notice that it was generative AI music?

386

And what do you think is the motivation for restaurants to use it?

387

I just want to say I think that I have subconsciously like repressed my awareness of that.

388

just this feeling of rage like came back up as you were saying that not towards you of

course but like towards this thing because yeah no aside from like being a product snob or

389

whatever anybody wants to think about me as like the thing that makes me the most angry is

like like generative AI songs in public spaces.

390

It really it's like a sonic infection.

391

It's like a it's like a total intrusion.

392

It's a...

393

Basically how I discovered it was happening was it's obvious because I just...

394

You can tell the difference.

395

It's like...

396

And then the second thing was I started using Shazam every time it would happen.

397

And one of two things would happen either it doesn't show up, and if it does show up it's

like the artist's name, the album name, and the publisher...

398

we're all the same title and be some random thing and there'd be some AI artwork and you

just realize like what's going on.

399

I started going up to the baristas at the cafes and the cashier at the front of the hotel

lobby.

400

mean this is happening in this has happened in three hotels, five cafes, three or four

restaurants now.

401

I mean like almost a dozen experience of this and every time when I go up I'm like, Hey,

just had a curiosity like

402

where's this music coming from?

403

And they're like, let me check.

404

And they're like, they show me their phone and it's just every time it's a YouTube

playlist and it'll be like chill lo-fi cafe beats or whatever.

405

And it's just a 14 hour playlist.

406

Yeah.

407

And so what's happening is some black hat SEO marketing type people realized that they can

408

put out these long form playlists which solve a business problem.

409

Think about the problem that solves for businesses.

410

First of all, if you're a cafe, you're like, and you're by the way, if you're a cafe in a

place like Thailand or Vietnam where, you know, they don't want to be playing traditional

411

Vietnamese or Thai music because they're supposed to be hipster.

412

And so who's defining like what a hipster cafe is like?

413

Well, the West and Europe in a way.

414

I mean, obviously they have their own amazing drinks and styles and in many.

415

ways aesthetically, musically, they go for all the same stuff that you get from your run

of the mill cafe in Portland, Oregon, which is like, it's it's chill lo-fi beats or

416

something like that, And so they're just kind of emulating that, but they don't

necessarily come in the same way.

417

Having worked in cafes, there's a certain pride you take of getting to take control over

the auxiliary cord and plug into your own.

418

you know, collection of music that you've curated.

419

It's like the one perk of working at a cafe is you get to put on your own playlist.

420

And so maybe there's that part of it that bothers me of like the cultural degradation of

like, cafes being like that third place that's not work or not home that you can go to and

421

which should be like a place where culture, I mean, traditionally, right?

422

Like cafes are where you exchange ideas and like, or you work independently, but you hear

something really stimulating that like,

423

Shazam and it's like, I've never heard of this artist before.

424

It's like the barista.

425

They look cool and they make great coffee and they've got great taste in music.

426

Like that's the stereotype and the great music is an important part of the experience, the

experience economy, let's call it.

427

So I just think that like for me having, you know, listened to a lot of Suno and UDO music

and generated a lot of it, it is that first of all, you'll see one of my posts if you go

428

back a month or two.

429

was about the, somebody did a data cloud analysis of Suno and UDO, I think it was Suno

specifically, and it was like the concentration of these keywords, and it was like

430

whispers, it was like all these like whispers, hearts, like it was like 150 words, and I'm

like, those are all the words, like those are the words you hear that are dead giveaways.

431

The second is like this kind of, it's like not quite auto-tuned and like the cool like,

432

R &B trap sense, but it's auto-tuned like every it's like if somebody played a vocal

melody on a synthesizer and then like also the arpeggiation of like a piano or a guitar

433

part it'll be like but it did it it it like really clean like too clean It's all these

things combined.

434

You just hear and you instantly know it The same way you're like, how do know that's a

Beatles song?

435

It's like I don't know because it's clearly line in a McCartney.

436

It's their songwriting It's their instrumentation is George Martin.

437

It's it's them, you know

438

Anyway, so between Shazam, looking at the YouTube, it's like piecing it all together.

439

And my girlfriend, by the way, she can recognize it.

440

And she's not even a musician.

441

But we walked into a place and she looked at me she's like, Suno, right?

442

And I was like, yeah.

443

So I don't think this is like, well, he's just a trained musician.

444

I don't think that's it.

445

I think what it is is that in Asia, they don't even have the conception of Suno and

generative AI music because these are baristas and they're not

446

Thailand and Vietnam, it's not necessarily third world countries, especially in the cities

where this is happening.

447

We're talking about in Da Lat, in Da Nang, in Ho Chi Minh, in Bangkok, in Chiang Mai.

448

These are the tourist hubs of these places, especially for Westerners.

449

But I don't think that the average barista there knows what AI music sounds like.

450

And so they just don't even realize when they're putting these playlists on what's

happening.

451

My second more conspiratorial view with the hotels, this could be wrong too, maybe this is

a stretch, but I just also wonder if they're ducking performing rights organization fees,

452

like PROs, because when you make a certain threshold of money in the states at least, you

can get audited and there's a huge fine if you're caught playing music you haven't

453

licensed, which means that people pay to license music.

454

I have a quick anecdote for this, but when I was a teenager,

455

I used to work at an ice cream store called Cold Stone Creamery.

456

Do they have that in Norway?

457

It's so cold there, you guys probably don't even eat ice cream.

458

So Cold Stone is like a McDonald's for ice cream, you can think of.

459

And we had a box, we had a physical box that was mounted on the wall with like 14 song

playlist.

460

And had like Paul McCartney, Silly Love Songs.

461

They had like these like hit songs, you know?

462

and would play the same 14 tracks every day all year round.

463

Except for like Christmas and then they'd have like the Christmas period and whatever.

464

But they licensed these songs probably for unlimited use across all franchises.

465

And so like as an employee you have to just suffer the pain of the 14 same songs your

entire time you work at that company, right?

466

But that's how I realized like this is how PROs work is they like squeeze companies and

are like, you need to pay us all this money.

467

But if you're hotel, in Asia, you either are like, we don't even, we don't respect

intellectual property anyway, so we'll play whatever we want.

468

In which case then they're just playing, they're streaming music because why not?

469

Because it's the same reason the cafe is.

470

Or what happened to me was I stayed at one hotel and then I went across the street and

stayed at the other hotel and they were both playing Suno tracks around the clock in the

471

lobby, around the clock.

472

Same type too, just like pop, Suno pop music.

473

I used to work for Suno.

474

I'm not, this isn't just rallying against Suno because people love to hate on them, but

it's about music that the lyrics were generated by chat GPT.

475

The music was clearly just like a simple prompt that they just generated a thousand

iterations of dumped into a YouTube file and uploaded.

476

is a low effort, low return kind of equation.

477

And for that to then find its way into public spaces, I think is actually really, I would

even say dangerous.

478

I think it's dangerous for human culture long-term.

479

I've actually heard of instances of this in Norway as well, because as you mentioned in

Thailand they might not even be aware of what's the genuine authentic music as compared to

480

the AI generated ones.

481

I heard of this at a cafe in Norway as well where a Norwegian jazz saxophonist was eating

and he was just noticing

482

this awful generic jazz music playing there.

483

And in Norway cafes pay kind of this blank slate rights payment either way.

484

So they don't have a financial incentive to use generated music because they have to pay

the rights organization anyway.

485

But this saxophonist then asked the cafe owner.

486

What's this music?

487

And then similar to what you were saying, the cafe owner showed a list of like 14 hour

nonstop generic jazz music.

488

And he was like, why are you playing this and not genuine music?

489

And because of the cafe owner probably doesn't even recognize the difference, but the

guests do.

490

And the reason the cafe owner had for using that playlist was simply because it was

491

very long non-stop, non-commercial break music.

492

basically just logistics.

493

It was a bit simpler.

494

Convenience, yeah, that's the word.

495

So yeah, that's definitely something to combat, I think, to not dilute human culture and

quality.

496

Totally agree with you on that.

497

You did mention it briefly actually in your last statement regarding Suno using chat GPT

for lyrics and that was kind of one telltale sign of generative music that you could hear

498

these kind of buzzwords that often gets repeated in Suno.

499

Which brings me to another question.

500

Why does Suno's lyrics suck?

501

Yeah, well, chat GPT's lyrics suck.

502

And why would Suno use chat GPT?

503

And this is pure inference, by the way.

504

I have no idea.

505

I actually haven't even heard.

506

Suno or UDO officially come out and say that they use chat GPT.

507

It's just been inferred, not just by me, but by many others online as well.

508

It's pretty clear that that's what's happening, especially because if you type in like a

prompt on chat GPT and then you put the same thing into Suno and UDO, you get the same

509

cliches across all three.

510

Like you can see it clearly if you do the experiments.

511

I think this is just my pet theory that they're passing on

512

Well, okay, let me back up for a second.

513

Are you familiar with when Anthropic, believe, I hope I'm getting this right, but I think

it was Anthropic got in trouble because Claude had recreated certain copywritten lyrics in

514

their outputs.

515

Right, you have heard about it.

516

man, wish, I'm pretty sure this is correct, that it was UMG who went after Anthropic and

was suing them.

517

And basically the courts were like, well, their takeaway was that Anthropic needed to do a

better job policing their outputs and essentially like gatekeeping to make sure that

518

copyright material didn't come out of it.

519

UMG was like, we don't even want them training on our stuff.

520

And basically like the courts aren't ready to do that.

521

I think there's a whole like meta narrative here around like the legal system and like why

all these AI generative AI companies are basically getting off scotch free.

522

And I think it has to do with geopolitical, like technological arms races with China.

523

you know, there's like bigger stuff happening than just like artists livelihoods, you

know, it has to do with.

524

with technology and, but to get back to the point, like the reason that I think Suno and

UDO and probably Refusion are all using ChatGPT is they can pass the buck on to OpenAI.

525

I mean, basically OpenAI is taking all the heat.

526

Now, if they had trained on something like, I always thought why not train on Genius?

527

Do you know the website Genius?

528

It's like a, it's a lyric database, right?

529

And it has like on every lyric,

530

on every stanza, user generated comments on what they believe that lyric means.

531

It's like a Wikipedia for lyrics, right?

532

So why not scrape all that the same way they scraped everything else?

533

I mean, it's not like it's not clearly not a moral dilemma for them.

534

Why not scrape that and then take all the metadata about the meaning of those lyrics and

then the genre of music that it belongs to.

535

And when somebody says like, create a trap song with trap lyrics, why not use the

language, the lexicon of that genre?

536

Why not?

537

I mean, not only do you have the lyrics, but you actually know like what all the double

entendres and the slang means.

538

It's all there in Genius, so why not do that?

539

Well, I think because if they were to do that, first of all, they could get sued by a

fellow tech company.

540

I mean, remember like they're an A16Z company, I believe, Genius.

541

one of the earlier ones, know, it's like these VC tech companies, the investors are

probably buddies.

542

So they would rather pass the buck to OpenAI, again, this is just my theory, and not have

to deal with stepping on the toes of a sort of cousin organization, let's call it.

543

I think that's why the lyrics are.

544

Yeah, so if there was any copyright issues with the lyrics in Udio or Suno, they could

just say that it wasn't us, it was ChatGPT, so they have to take the fall for it.

545

Yeah, and it totally makes sense because as you say, why not train it on, why not scrape

the data from Genius that would obviously be a lot better than what we're currently using.

546

Okay, so.

547

Let's take a slight pivot here.

548

I know that you have done some educating using hook theory and doing music theory lessons

for music producers, more for producers maybe than the musicians.

549

So you have some experience there in using technology for kind of music technologists.

550

And as you were saying, AI is a huge disruption for

551

society as a whole and probably will be in education as well.

552

And I'm wondering if you have any take on if music education needs to change to adapt to

AI and if so how?

553

Hmm, and that's such a good question.

554

Yeah, I Well, I think there's like a few layers to this because first of all the industry

is changing which means career paths are changing which means career preparedness Isn't

555

the same thing that it was ten years ago, right?

556

So if you have if you know right today the tools the generative tools and the machine

learning tools that we have They may not truly be competitive.

557

But like how long until they are?

558

I don't know.

559

And what do any job markets look like when AI becomes like the standard hiring procedure?

560

Like if an AI can do it, it's cheaper to do that than hire a human, right?

561

So what does the industry look like?

562

So when you're talking about music education,

563

There's the layer of music theory, music production, this traditional skill set.

564

You could say, well, that's gonna be disrupted because instead of writing a melody from

scratch, you're gonna be procuring melodies from a generative system and assembling them

565

together like a Frankenstein.

566

Right?

567

Like a sample-based producer, really.

568

But instead of licensing samples and downloading sample packs, you're going to be

prompting and generating and fine-tuning and, you know, if we talk about specific tools,

569

there's like, you know, Samplab and RIPx, which are audio to MIDI transcription tools.

570

So you can take, you could generate something in Suno or UDO or Refusion, stem separate

it, grab the particular track, throw it into RIPx or Samplab, manipulate the individual

571

notes.

572

pass it through a plugin like Soothe 2, which is able to kind of take some of the high

ends off so it's not so scratchy.

573

And all of sudden, okay, you've actually got a decent sounding thing, run it through some,

Lander has an AI music plugin suite, so okay, run it through that for mastering.

574

you know, I mean, there's like all these different sort of, my friend, Dime Doogle, he's a

producer, production music composer, and...

575

He calls it like a modular AI stack.

576

And I think that's like a really great way to think about it.

577

Like the same way, if you were in Ableton and Max and you were like, you're creating like

a modular arrangement for effects, like effect chains or something.

578

You can imagine in a similar way doing that with AI tools where it's like, I generate it,

I stem separate, I clean it up, I do note manipulation, I do automatic EQing.

579

Maybe even there's some other interesting stuff that hasn't surfaced yet.

580

have some ideas.

581

what that could be, and then there's a mastering stage.

582

That's a very different thing from what music composition was 50 years ago, where it's a

guy or a girl with their guitar or piano and imagining a whole symphony.

583

these, now look at what the average community college in America is teaching is still Bach

584

It's like.

585

Hello.

586

It's so far off base from where things are at.

587

And I get that like, we need to teach music education, but I think that part of the way

music education has to adapt is to like square up with the reality of the world we live in

588

today.

589

That doesn't mean that we shouldn't be training skilled.

590

instrumentalists or composers who understand complex theory and harmonic analysis and can

really do that stuff manually.

591

Because I do think the role of humans in relationship to AI will be a curator and

arranger, maybe a little bit of production, but also there will be a need for fine tuning

592

and thinking about sonic brand.

593

So it'll be more about your imagination.

594

and your storytelling capacity as a musician and less about your physical capabilities.

595

I think that's the biggest disruption we're gonna see on the musician, composer, producer

side.

596

But when you talk about the industry, you talk about music supervisors, like roles like

that, like sync music licensing.

597

this is, mean, the majority of musicians are not making money, at least like they're not,

you know, serious money off of revenue from Spotify streams and Apple Music.

598

I mean, maybe if you're lucky, you're making a few thousand dollars a month.

599

If you're very lucky, you know, my band does 20, 30,000 monthly listeners and we do like

200 bucks a month.

600

So I mean, I don't know the fraction of 0.0002 percent per stream and based on how long

you listen, I mean, it's crazy.

601

So the money is in licensing.

602

It's in sync licensing.

603

You know, that's like, or if you're really fortunate, you get brought on to a team for

604

you know, for a video game, these were the custom composing for a score, but that's the

extreme minority.

605

I most musicians are not even aiming for that, you know, for film music and video games.

606

So I think that a lot of that stuff is going to be generative.

607

And I think, you know, education needs to prepare people mentally for the actual industry

that's incoming.

608

And that means being equipped to use the tools, which is why I do what I do and which is

why I don't like the hype machine because I think the hype machine gets in the way of

609

actual industry awareness.

610

I mean, if we're all just hyping each other's stuff because it has the word AI in front of

it, and we're resharing it we're not even...

611

taking a moment to understand it, then how can we educate ourselves and how can we educate

others if we're teachers, you know, when we don't even understand the tools?

612

So, I mean, I think it's great that you're actually testing them, it sounds like yourself

and you're having the same experience I am, which is like, wow, the proposition, the value

613

proposition of the app versus what it actually does, there's still like this gap.

614

Music is the worst one.

615

I mean, video, AI video and AI image and AI text is like...

616

leagues beyond where music is right now.

617

Suno and Udio were like, hey, just so you guys know, it could be way better.

618

We did something horrible, but at least you know that it could be better.

619

Until then, we had Boomy and Hubert and Ava and these first, what I call the first wave

tools.

620

People were like, well, I can't really do it because music is a spiritual human blah,

blah, It's like, sorry.

621

Music is a quantifier, it's just like image and video and everything else.

622

That doesn't mean that the human spirit doesn't matter.

623

It's incredibly important.

624

It's the reason that when we listen to an AI generated Suno song in a cafe, it's

disgusting.

625

It's disgusting because artists weren't compensated, the music wasn't licensed.

626

It's disgusting aesthetically because it is not creative.

627

And it's disappointing that there's no regulation, that the streaming platforms, even

though we have tools like IRCAM Amplify, which is I think one of the best AI content

628

detection tools that I've tested, I've tried a couple other ones, Pex, there's a few of

them out there.

629

You know, they get it to a certain degree of accuracy, but I think the problem is they

can't do 100 % accuracy with content detection because the algorithms, the generative

630

algorithms change so often.

631

so, you know, anyway, music education is gonna have to account for, you know, artists'

livelihoods, copyright, the redefinition of copyright law, what it means to make music

632

composing as a profession, all this stuff.

633

And...

634

You I'll close with this.

635

say, you know, one of my clients, future proof music school, they're actually

636

former head of education from Icon Collective, John Von Siegern and a couple of people

from that team have branched off and they've created, they have human mentors, human

637

created courses, but they have AI that asks you questions and learns where your knowledge

gaps are and creates a custom curriculum around that.

638

And they have an AI chatbot that you can upload music tracks to and you can get real

feedback while you're in the DAW.

639

It's actually voice activated.

640

So as I'm speaking to

641

Cadence is the name of the bot.

642

Cadence is like, have you tried adjusting that parameter?

643

I'm like, where is that on the interface?

644

They're like, try looking in this there.

645

I'm like, I don't see it there.

646

And they're like, what about this?

647

It's as if you have a mentor.

648

like speaking to you in real time, but it's available 24 seven.

649

And so for me, I found that's like the most profound AI music education tool I've

personally used because like it was just in real time giving me the kind of support I

650

usually have to pay like a premium for.

651

And yet there's certain things it doesn't understand, it can't see, and it can only hear

with a certain amount of intelligence.

652

Like it has limits to what it does and doesn't understand.

653

So that's where then you can.

654

you can level up and get that high touch mentorship call like a couple times a month if

you need it.

655

And there's a discord community.

656

And so it's like you have your peers and your teachers and your AI all together in one

spot.

657

I think that's the future of music education is like these AI augmented curriculums.

658

That sounds cool.

659

I have to try that out, I think.

660

The school I work at is kind of known for electronic music.

661

yeah, by the way, I don't work for the classical department.

662

So I work for what's in Norway called the Rhythmical Department, which is internationally

known as popular music.

663

And so because of this institute's devotion to

664

I guess you could say modern music interpretation.

665

That's why we've been awarded with this center of excellence as it's called, where we work

on the creative use of technologies in music education.

666

And when I discuss with other faculty members, particularly the ones doing electronic

music, the things we tend to focus on are risk taking and collaboration.

667

So that's kind of some directions that kind of, or at least hopefully will be relevant no

matter what kind of technology comes.

668

Because in some way this AI wave is kind of the end tale of a long development of MIDI and

synthesizers and drum machines.

669

And there's been a lot of stuff that had the potential of automating

670

musicians' tasks, but we haven't been fully automated yet and I suspect we will never be

unless we reach this Westworld or Matrix dystopian future.

671

But yeah, we'll get there.

672

No worries.

673

you on the risk taking side of that, can you expand on that a little bit?

674

What do you mean by that?

675

So that could mean playing live for instance, in contrast to just sitting at home

producing stuff.

676

And also there's a risk taking in involving other people in your project.

677

And it could be a risk in terms of using plugins that you don't exactly know how it will

behave.

678

Stuff like that.

679

recordings of your foot splashing in a puddle of water that your teacher randomly poured

on the floor.

680

Yeah, I like that.

681

You know, yeah, that's really beautiful.

682

I think that those are like holistic, universal truths about music making, which is like

the more risks you're willing to take.

683

Yeah, you may fail along the way, but.

684

the more likely you are to stumble on something interesting.

685

And I think that when you work within the constraints of something risky or unusual,

there's this higher probability of stumbling across something kind of cool.

686

Yeah.

687

There's a lot of art in mistakes.

688

I have this professor at the university who has this quote regarding mistakes and AI and

language models, particularly concerning the Norwegian band Aha.

689

They had this huge hit called Take On Me.

690

I don't know if you've heard that probably.

691

Yeah.

692

Take On Me, Take Me On.

693

It's not very linguistically correct that song.

694

So I mean, they might never have made such a hit if they did have ChatGPT to correct their

grammar.

695

So there's some beautiful stuff lying in the mistakes.

696

So we're starting to get to a close here, Edra, but I was wondering if you could...

697

if you could say something about what excites you most right now when it comes to AI music

tools.

698

Well, I'll tell you the dream tool that doesn't exist, and I wish that it did.

699

How about that?

700

Maybe somebody will see this and they'll be like, I'm gonna build that.

701

And then they can send me the beta and I'll trash it on LinkedIn.

702

And then eventually it'll get better and better until I'm supporting them on LinkedIn.

703

I would love to see a tool that could infer the timbre of a particular instrument from a

track.

704

So for example, like,

705

I would love to be able to say, me the piano from this track.

706

Like, and let me just play that on the MIDI keyboard.

707

The amount of time that I spend as a composer trying to just get a sound that doesn't

suck.

708

Think about, and I know this would be horrible for Native Instruments and all these

companies where it's like, they just sell thousands and thousands of dollars of product

709

that are really high quality, incredible, all the different articulations of a piano from

pianissimo all the way to forte, you know what I mean?

710

But it would be amazing.

711

to be able to get something like that where you can infer.

712

There is actually a company called Samplab that I mentioned earlier, there's audio to

MIDI.

713

I'm beta testing a tool right now for them called Reharmonizer.

714

And it is kind of this idea, but it's like just a one shot.

715

So it can analyze a one shot and it turns it into a wavetable and it tries to recreate

that.

716

It's like a little bit noisy because it's a generative recreation and then it pitch shifts

it.

717

It does a pretty good job, but it's like, I haven't seen anything that really

718

like does this yet.

719

There's synth GPT where can type in some text and turn it into what you're going.

720

But again, like it's doing that with a wavetable.

721

The sounds are a little generic,

722

These are all proofs of concept.

723

We're in the proof of concept phase of AI music plugins in general.

724

We just have to be okay with that and come to terms with it.

725

Like none of these tools, mean, even AudioCypher has so many problems.

726

I it's not buggy, but the problem I have with it is that I would like to be able to toggle

over from text to MIDI, melody and chord generation based on cryptograms to a tool that

727

actually generates MIDI that feels like the words that I put in.

728

say like the feeling of standing outside with your bare feet on a snowy day and it just

creates a composition that actually evokes that feeling.

729

To do that it also needs a sound design layer.

730

like.

731

because MIDI on its own, the sound design is everything actually.

732

It's not everything, but it's certainly at least 50 % of it.

733

So my dream for audio cipher would be to get millions of dollars from a VC firm who just

sees the vision and just build out the actual product I have in mind, which is turn your

734

ideas into music, but not finish audio samples, content in your DAW that you have total

control over, rather than what we have now, which is like stem separate a Suno track.

735

then pitch shift the individual notes and then it sounds all grainy and messed up.

736

It's like that's fine for now, but ultimately I want to have tools for composers and

musicians in the DAW.

737

That's my dream.

738

Yeah, awesome.

739

So thanks so much for joining Ezra.

740

It's been a pleasure.

741

Yeah, likewise, Don, it really nice to meet you and talk to you about this stuff.