Is it AI or is it Memorex?

I have a confession: I’m an AI-music skeptic. There, I said it. It might seem strange because I'm also the co-founder of ever.fm which looks, at first glance, like an AI project. It has randomness, it uses computers, it lives on the internet. So, what gives?

I remember meeting Alex Mitchell of Boomy.com in 2019, before AI was mainstream news. He said something that really stuck with me: "If a project involves computers and randomness, people will call it AI." The term "AI" is a bit like the term "natural", an unregulated word that gets slapped on a lot of products. What does "natural shampoo" or "natural flavour" really mean? It's murky at best. So, too, with AI.

You could say that at ever.fm, we do use AI. But for us it means "Artist Intelligence." That's a very Canadian way of saying that when it comes to generating artful music, generative AI alone isn't up to the task. It's nothing personal. And that's just it. It isn't personal.

When it comes to generating artful music, generative AI alone isn't up to the task. It's nothing personal. And that's just it. It isn't personal.

Now, before I get dragged by my AI maximalist friends for saying that, let me add two more things:

1. Generative AI can make beautiful things. But beauty is not art.

2. Generative AI might make art someday. But it would mean re-defining what art is. And that takes time. We're not there yet.

ever.fm works with 100% artist-created audio. We are a post-production phase, not a replacement for recording or production software. By the time artists are using ever.fm, they have already created all the audio ingredients they need. They bring those ingredients (audio layers called "stems") to us in order to animate the arrangement. In fact, it's very similar to cartoon animation: first you design the character as a digital "puppet", then you animate it. Wow, as I write this I'm realizing how analogous that is to ever.fm. We are animating audio, taking an unchanging mix of a song and giving it movement, variation and personality. We're bringing the mix to life.

We are animating audio, taking an unchanging mix of a song and giving it movement, variation and personality. We're bringing the mix to life.

In our modern media landscape it's easy to get jaded and jump to conclusions. When you first hear about ever.fm you might immediately assume it's an AI-driven gimmick. “Infinite versions of one song?” Yeah, okay. Probably just some fancy diffusion model spitting out endless remixes based on listener data or something. But take a closer look (and listen) and you'll hear the decisions we made at the outset of this project, the DNA of ever.fm that puts human-created music at the centre. Yes, there's an element of chance—you don't know exactly how the chips will fall with each shuffle. But that element of chance is part of the composition itself. It's a feature, not a bug. It's less John Connor and more John Cage.

There's no Artificial Intelligence here, just Artist Intelligence. Every sound is recorded by the artist, and every rule for how the song changes—from the instrumentation to the structure to the overall vibe—is set by the artist. The artist's choices are extension of their composition. And if you had pegged ever.fm as yet another example of tech encroaching on art, think of it more like tech collaborating with art. Instead of replacing musicians with machines, ever.fm gives artists a way to expand their creative vision in a meaningful way, bridging the gap between the controlled environment of the recording studio and the energy and unpredictability of a live performance.

There's no Artificial Intelligence here, just Artist Intelligence.

This isn’t about machines taking over music; it’s about artists exploring the full potential of a song. Think of it like this: in a traditional release, an artist picks one version of a track, polishes it until it shines, and says, “Here’s the song, take it or leave it.” But with ever.fm, they get to say, “Here’s the song, but here’s also what it could be.”

It’s still the artist’s voice, amplified. It's the artist's vision, expanded. Listeners get to engage with the song in a new way and by participating in the curation of the music, fans go behind the curtain. They get to hear the multitudes contained with a single song.

There’s no uncanny valley here, no synthetic weirdness masquerading as creativity. Just an artist playing with the infinite possibilities of their own work and dialoguing through art with anyone who wants to lean in.