Warner Music signed an algorithm to a record deal

Last week, a press release went out to tech and music reporters claiming that little-known startup Endel had become the “first-ever algorithm to sign [a] major label deal” with Warner Music.

The news was covered widely, with commentators tossing around phrases like “the end is nigh[1]” while hand-wringing over the idea of coders coming for musicians’ label contracts. But the press release wasn’t exactly right, and questions about the future of music are even bigger than anyone thought.

Endel is an app[2] that generates reactive, personalized “soundscapes” to promote things like focus or relaxation. It takes in data like your location, time, and the weather to create these soundscapes, and the result is not quite “musical” in the traditional sense. It’s ambient, layering in things like washed-out white noise and long string notes. It’s the type of stuff that’s exploded on streaming platforms in recent years under newly invented genre names like “sleep.”

Although Endel signed a deal with Warner, the deal is crucially not for “an algorithm,” and Warner is not in control of Endel’s product. The label approached Endel with a distribution deal and Endel used its algorithm to create 600 short tracks on 20 albums that were then put on streaming services, returning a 50 / 50 royalty split to Endel. A publishing deal is unlike a typical major label record deal in many ways. Endel didn’t get any advance money paid upfront, and it retained ownership of the master recordings.

Even if Endel had signed over the masters, the company could easily just make more: Dmitry Evgrafov, Endel’s composer and head of sound design, says all 600 tracks were made “with a click of a button.” There was minimal human involvement outside of chopping up the audio and mastering it for streaming. Endel even hired a third-party company to write the track titles. Five Endel albums have already been released, and 15 more are coming this year — all of which will be generated by code. In the future, Endel will be able to make infinite ambient tracks.

It’s not entirely shocking that the type of audio Endel creates generated interest from labels like Warner Music, which has a history of investing in music AI products. While the top playlists on Spotify are made up of traditional genres like pop and hip-hop, streaming services have a bunch of weird categories that have recently become very popular. These generally fall under the kind of things you have on in the background, like “lo-fi” or “ambient.” They’ve become popular due to the rise of what’s called a “context playlist[3],” which is a playlist that doesn’t cater to a genre but instead targets events or moods like “Peaceful Piano” and “Deep Focus.” These playlists have millions of subscribers and are easy wins for labels like Warner — especially if they can be filled cheaply.

These playlists briefly became a controversy in 2017 when there was a conspiracy theory about Spotify astroturfing chunks of chill and piano-driven playlists with “fake” artists[4]. In the end, it turned out to be a nonstory, as the “fake” artists were aliases for real musicians. Many of these artists are composers with no interest in becoming famous artists who sell ownership of their recordings to library content companies. The companies traditionally just provided royalty-free soundtracks for videos and podcasts, but they discovered a new revenue line from context playlists.

It started a debate that rages on: what is a “fake” song? If a song is created by people using aliases with no other online presence, does that make them fake? Should streaming platforms disclose when songs are made by “fake” artists?

If those questions were hard to answer two years ago, they are even harder to answer now that Warner has signed a deal to market and distribute 20 albums’ worth of material made by software. Streaming platforms like Spotify created a new market for context playlists full of sleep and ambient music, and music software like Endel might disrupt that market by lowering the cost of production to zero. But the industry hasn’t yet been thinking about AI that way.

In our Future of Music episode on AI last year[5], I talked with Michael Hobe, founder of AI platform Amper. “We can facilitate your creative process to cut a lot of the bullshit elements of it,” Hobe said. “For me, it’s allowing more people to be creative and then allowing the people who already have some of these creative aspects to really further themselves.”

Personally, I’ve tried almost every algorithm and AI-driven software package for music creation, from Amper to Captain Plugins[6] to Google Magenta Studio[7] and even the beta of Sony’s Flow Machines in plug-in format[8]. They’re all… fine. None of them are even close to replacing a Beyoncé hit on the radio. But they can be tools to give me inspiration or allow me to not have to worry about certain engineering bits so I can focus on the creative parts. As I talk to artists around the industry, the ability to cut out those “bullshit elements” seems to be driving the adoption of AI.

1 2