Mikey Shulman | Training Data Summary

Training Data Podcast | Participants: Mikey Shulman (CEO, Suno), Sonya Huang

1. Key Themes

AI as Creative Democratization, Not Automation

Suno's core thesis is that AI's most profound impact isn't replacing existing workflows — it's unlocking entirely new forms of human expression that were previously inaccessible. 90% of Suno's daily users create something, flipping the historic ratio of creators to consumers.

"Before Suno, basically everybody was a consumer of music. You know, compared to the 8 billion people on the planet, there are very few people who make music and the rest of us consume it... The crazy thing about Suno is that on any given day, 90% of the users are going to create something." 00:16:51

"We are much more inspired and motivated by doing something that wasn't possible until today instead of automating or speeding up something that already exists." 00:34:39

Music as Active Entertainment, Not Passive Consumption — The Gaming Parallel

Mikey explicitly repositions Suno away from the Spotify paradigm toward the gaming paradigm: engaging, brain-activating, fulfilling, and monetizable in fundamentally different ways. This is a strategic frame that redefines who Suno competes with.

"I will get shit inside of the music industry for saying that there's a lot to learn from gaming. But there's a lot to learn from gaming. And it's not just about how well it monetizes. It is how it grabs your attention and how it pulls you in and makes you use your brain. And there's also things about the business model about gaming that music has to learn from." 00:18:47

"The creation is actually the entertaining bit... It's like everybody in the world is creative. Being creative makes you feel a certain way. This is like in our DNA." 00:17:20

The Model Layer Is Necessary But Not Sufficient — Product Moat Is the Real Game

Mikey is unusually candid about the model arms race and why product and UX are the durable differentiator, especially against well-capitalized incumbents like Google.

"It's unclear how much moat exists in only a model. And it's like, I'll just say it. Like Google has started to build music models. And while ours are way better today, they're Google and they'll outspend us seven days a week. And they can probably catch up on the model side. And so I think it's just really undervalued to invest in the product and the UI and the UX to make sure that you're constantly delighting people." 00:28:26

2. Contrarian Perspectives

Music Is NOT a Scaling Problem — Smaller Models Win

Against the dominant LLM narrative that scale solves everything, Mikey argues music AI is fundamentally different. Smaller models are actually advantageous — they respond faster (critical UX) and the challenge is alignment to messy human taste, not benchmark climbing.

"Music is really not a scale problem. The models are pretty small for a variety of reasons... In LLM land, there's all of these benchmarks... scale is actually a pretty efficient way to climb up the ladder... In music, there are no right answers. There are no benchmarks. And so scale is somewhat less helpful in solving it." 00:12:24

"The models not being that big actually lets us get you the music quicker, which turns out to be really important for good UX." 00:13:20

Injecting Musical Knowledge Into the Model Is a Trap, Not a Shortcut

The obvious engineering approach — encoding music theory into the model — is precisely what limits the output. Suno's breakthrough was treating music purely as raw waveform data, enabling genre-bending outputs that couldn't exist within the constraints of human-defined music theory.

"The more musical knowledge we give the model, the more constrained it will be in a bad way... In Western music, there are 12 tones. If you tell the model there are 12 tones, it will only ever produce those 12 tones. You will be forever limited... let's throw away everything we know about music." 00:07:13

The Record Labels Are NOT Cooked — They're Strategic Partners

Contrary to the default Silicon Valley narrative that AI disrupts and displaces labels, Mikey argues labels are among the most culturally important institutions and the future is co-creation, not displacement.

"People also expect me to say like, oh, the record labels are cooked. I think that's obviously wrong. They're some of the most culturally important institutions in the world. They understand music and they understand music culture. They cultivate and grow stars that resonate with billions of people." 00:24:52

Full Songs Over Crisp Audio — The Contrarian Product Bet That Paid Off

While competitors optimized for audio quality (diffusion models, clean 10-second clips), Suno deliberately chose worse-sounding but narratively complete songs. The story beat the spec sheet.

"For the longest time, our audio is really not crisp. And every single one of our competitors had just way crisper audio... And to just go all in on that and say, like, okay, we're going to make full songs. And yes, they're not going to sound amazing. But they are still going to tell the story instead of making perfectly sounding audio that just is like background music." 00:30:46

3. Companies Identified

Suno AI music generation platform; described as the largest AI music company in the world, at ~$300M revenue run rate. Mentioned as defining the "creative entertainment" category — turning passive listeners into active creators. Notable milestone: Warner Music settlement and partnership, artists signed to record deals via Suno-generated music, chart-topping singles.

"On any given day, 90% of the users are going to create something." 00:17:20

Kensho Technologies Quantitative/AI analytics firm (acquired by S&P Global), where Mikey and several co-founders met. Cited as an exceptional example of talent density creating a lasting network of founders.

"The Kensho mafia is like pretty unparalleled... I just credit Daniel with that, honestly. Daniel is like, I think, the best object lesson in what talent density can do for a company." 00:05:44

Warner Music Group Major record label. Mentioned for being the first major label to reach a landmark settlement and partnership with Suno, signaling a shift from adversarial to collaborative in AI music.

"What I'm most excited about doing with Werner is actually building things together that could never have existed before and building products that let fans interact with their favorite artists and really deepen the artist-fan connection." 00:25:45

Midjourney AI image generation platform. Cited as the product inspiration for Suno's go-to-market strategy: Discord-first, low friction, community-driven validation.

"We kind of took the example of Mid Journey and we said, it's really easy to put a Discord bot out and see will people enjoy it." 00:04:19

4. People Identified

Daniel Nadler Founder of Kensho Technologies. Cited specifically for his exceptional talent in identifying and recruiting non-traditional, high-potential people — used as the gold standard example of talent density as a company-building strategy.

"Daniel is like, I think, the best object lesson in what talent density can do for a company. And it was a lot of people with non-traditional backgrounds. It skewed very young. But he was great at finding people and great at convincing them to join." 00:05:44

Harrison Chase Co-founder of LangChain. Mentioned as a mutual connection and early Suno Discord user, as well as a Kensho alumnus. Notable because his enthusiastic early adoption was a key signal validating Suno's product before it had a web app.

"Our mutual friends, Harrison Chase, was one of the earliest Suno Discord users. And he was having far too much fun making songs in your Discord." 00:02:14

Zania Monet Poet-turned-musician who used Suno to convert a decade of written poetry into music, found a new audience and artistic voice. Cited as Mikey's favorite example of Suno raising the creative ceiling — new artists with genuinely personal stories finding new channels.

"My favorite example is Zania Monet, who it's the stage name of a poet who took all of her beautiful poetry that she had been writing for like a decade and started to make music out of it. And found an entirely new voice and an entirely new audience to resonate with her art." 00:22:56

5. Operating Insights

Release on a Steady Cadence, Not Perfection Cycles

Suno deliberately avoids the "wait for the perfect model" trap. They ship on a cadence and define version cutoffs somewhat arbitrarily — preventing the organizational paralysis of waiting for a transformational release while maintaining momentum and user trust.

"What you would hate to have happen is we don't release stuff for like two years and we try to make, you know, the music model to save humanity... Almost just to keep it on a steady cadence of when we release things." 00:11:26

Human Preference Data at Scale Is Both a Research Asset and a Moat — Not Just a Training Input

Preference data doesn't just improve the model; at Suno's scale, it enables research techniques that wouldn't otherwise be discoverable. This is a non-obvious virtuous cycle: more users → more preference data → unlocks new research methods → better model → more users.

"A really underappreciated thing is how much this preference data actually lets us do research. Like without the scale of preference data that we have, we wouldn't even be able to develop the techniques that we are using." 00:13:20

Use the "Staying Up Late" Test as a Go/No-Go Signal on Product Direction

Mikey's heuristic for committing to a product direction: if the team is voluntarily staying up late playing with the thing, that's a stronger signal than any market analysis.

"When you are staying up late playing with the thing and you don't want to go to sleep, it's like a really good sign that that is what you are meant to be doing." 00:04:48

6. Overlooked Insights

Voice Personalization May Be the Killer Engagement and Retention Feature Across All AI Creative Platforms

Mikey mentions almost in passing that Suno's most recent feature — using your own voice in a song — creates a disproportionately strong emotional attachment, and dramatically increases the resonance of songs you send to others. This is a product insight with implications far beyond music: identity-embedded AI output is stickier and more viral than generic AI output.

"When you hear yourself in the song, you get so much more attached to it. But actually even more so is when I send you a song and you can hear me in it, that song will resonate much more than some nondescript voice, even if that nondescript voice is very good. And it's because the human ear is highly attuned to voices." 00:32:35

This suggests any AI creative platform that can embed personal identity (voice, style, likeness) into the output will see step-change improvements in retention and sharing — a blueprint for product differentiation in AI consumer apps broadly.

The "Professional Music Has Suno In It Already" Disclosure Is a Major Signal for B2B/Pro Tier

Mikey drops almost as an aside that charting professional tracks already contain Suno components — not fully AI, but as a workflow tool. This points to a quietly emerging and highly monetizable professional/prosumer market that Suno hasn't yet formally addressed but is already winning organically.

"I just know that there are tons of charting tracks that have little bits of Suno in it. They're not entirely Suno. And it's because for the professionals, it's also just an amazing tool to use as part of your workflow." 00:23:51

This is the equivalent of discovering your consumer product already has organic enterprise adoption — a classic signal to build a dedicated pro tier before a competitor formalizes it first.