Making Music in the Age of Generative AI

A couple of weeks ago, the CEO of the AI startup Suno, Mikey Schulman, went on the 20VC podcast and said some things about making music. For those of you who aren't familiar, Suno is that app that allows you to type in a prompt that specifies lyrics, subject matter, and/or genre and in just a few seconds it'll produce mp3 audio of a fully formed song, to your specifications.

So anyway, on this podcast, Schulman said:

“It’s not really enjoyable to make music now… it takes a lot of time, it takes a lot of practice, you have to get really good at an instrument or really good at a piece of production software. I think the majority of people don’t enjoy the majority of time they spend making music.”

He later went back to amend/expand on this statement in a tweet:

This quote makes it sound like I'm discouraging people from learning music, enjoying music, or improving their craft — I didn’t mean to ... there are two unfortunate and common cases where people miss out on the joys of making music: (a) people who drop their craft out of frustration before they get good enough to truly enjoy it and (b) people who struggle through, learn an instrument to a high level, and then drop it because they couldn’t keep up with the practice or it became stressful. How many people do you know who stopped playing their instrument when they left high school?

Let me first say that I don't know Schulman personally nor do I have any reason to believe anything he's doing is malicious or intended to harm anyone. I'm happy to assume he's doing his best to lead a company built around some really cool tech and creating meaning around that effort.

That being said, unsurprisingly, I have thoughts! There is a lot to unpack here and I could literally talk about this all day, but I will try to distill my thoughts down to just a few points.

Using Suno ≠ Making Music

Let's be clear: using an app like Suno is not "making music." The idea that using Suno (or tech like it) is making music is akin to the idea that using a calculator is the same as doing math. You didn't do math, the calculator did. It'd be like using Midjourney to create an image and saying that you "took a photo."

"But people use technology to make music all the time!" Yes, absolutely. From the electromagnetic pickups that allow guitar sounds to be pumped through amplifiers to programming virtual instruments in Logic Pro, musicians have been using technology of one kind or another to make music for decades (if not longer, depending on how you define technology).

The thing is, with Suno, the human is no longer making music. I would go so far as to say Suno isn't even making music. It's producing audio content that sounds a lot like music as we know it.

In the Hayao Miyazaki film Spirited Away, the character No-Face tries to offer the protagonist, Chihiro, handfuls of gold because he thinks that's what Chihiro wants. She doesn't want the gold and, as it turns out, it wasn't even real gold, just rocks and dirt dressed up to look like gold.

This is what AI tools like Suno are doing. They are confidently offering up what they think the user wants to hear, but the result is not real music. It's audio content dressed up to sound like music. It's a simulation. Fake sugar.

It's not about using tech to make music. It's about letting tech make music for you. Music, like all art, is a human expression of the human experience. It is meaningful to us as humans because another human is communicating their ideas to us through their art. AI, by definition, is not human, so it cannot produce art. Therefore, Suno is not making music.

Learning is frustrating

In the podcast interview, Schulman talks about how it's not enjoyable to make music anymore. In his amended statements, he refers to people who couldn't get over the learning curve of music and stopped out of frustration because they weren't "good enough to truly enjoy it." The implication is that Suno would address this issue.

The problem here isn't frustration. Frustration is a natural, normal part of learning. In fact, productive struggling is key to learning. The problem is ceasing to enjoy being frustrated. The cause is when good enough isn't "good enough."

Here's the thing: making music is inherently joyful. Put a guitar or piano or drum in front of a child who has never played music and watch their face when they play a few notes. You'll see delight, elation, intrigue, focus... but not frustration.

So why do we get frustrated? Frustration generally comes from our brain wanting something or someone to be where it isn't. In music, we get frustrated when the thing we're playing does not match what we think we should be hearing. In other words, when we are learning, frustration is inevitable because we are pushing ourselves to do something we currently cannot do.

The root of the kind of frustration that causes people to stop learning is a misinterpretation of what is "good enough." This usually occurs when our expectations are not realistic. If someone comes into their first voice lesson expecting to sound like Beyoncé after an hour, they're going to leave frustrated and likely not come back. On the other hand, if someone is coached properly, they will leave their first lesson with a realistic understanding of how to begin improving. Will they ever sound like Beyoncé? Maybe. Will they be frustrated? With any luck, yes! But if they are taking steps towards achievable goals, they're far more likely to enjoy the process and come back.

All this to say, a shortcut, a "button to press" that skips all of the frustration of learning, a tool that spits out a song from just a few words of descriptive text, will not make learning music more enjoyable because it is circumventing the very mechanism (struggle) that makes learning productive, and therefore enjoyable, in the first place.

Craft is lifelong

Schulman makes reference to people who stop playing after high school because of the stresses of practicing. Unless you are a professional musician, the intensity and rigor of your practice regimen is self-imposed (meaning, your practicing does not impact your ability to eat or pay the rent).

If someone played at a "high level" (whatever that may be) in high school then stopped playing music altogether, I would first wonder why they were playing at such a high level to begin with. Did they do it for enjoyment? Did they do it to satisfy their parents? To impress a romantic interest? As a resume piece to get into college? If the answer was not primarily "because I like it", I wouldn't be surprised if people stopped playing. But I highly doubt it would be due to the practice regimen.

This brings me back to the bit about expectations. If your expectation is that playing music will help you get into college, then you get into college, then of course you'll have little motivation to continue playing after high school (regardless of whether you get into college or not). If, however, your expectation is that you will enjoy playing just for the sake of playing, then it won't matter what happens in college. Maybe you'll even stop playing for a while... but eventually, it will come back.

I always say that I love having adult students at our school because they have the best stories. They always have some wonderful reason for taking lessons. They want to do something for themselves. They want to surprise their children. They want to meet new people. Some are starting from scratch because they've always wanted to learn but never had the opportunity to do so. But most folks have played before and want to get back to it. It is never too late to learn (or return to) music!

I love technology

In closing, let me be clear: I love technology. I lived, breathed, and worked in tech for 15 years. I use technology every day to make my life and work easier and I am a firm believer that technology has done wonders for music (the Fender Play app and Ultimate Guitar are two off the top of my head that are pretty great).

I also believe that we are hurtling face-first into a time where the lines between technology and [insert human activity] are becoming exponentially more blurred. It's critical to have these discussions so that we are better prepared to handle what's coming at us.

As a former product manager, I empathize with Schulman. I don't fault him or any of his peers for building their companies and claiming that they can do X, Y, or Z. After all, they're just trying to find meaning in all of this just like we are.

Suno is a really cool piece of technology. I can't imagine the amount of creativity, thought, and effort that went into coding it, not to mention the work put into creating a company around it. Is there a place for Suno in this world? Surely. Is using Suno an enjoyable experience? Undoubtedly, yes! If you haven't tried it out yet, I highly suggest doing so.

That being said, we cannot conflate creating audio content and making music nor should we confuse frustration with dissatisfaction. In its current form, Suno, like many of its gen AI brethren, is a solution looking for a problem. When you're a hammer (even if you're a really cool, innovative hammer) everything looks like a nail.

For now, I will look forward to all the ways in which tech will continue to help make learning music more enjoyable.