Algorithms in music: is it the future?

Why do people buy records? Is it for the all the inconvenience of buying an audio setup, and the expense of always having to shell out $20 to $40 for each record? No. It’s for the general sentiment that comes from physically owning music, and being forced to listen to an album all the way through; no mash-ups or skipping tracks. But what happens to that sentimentality when most people just listen to music digitally, on their phones, with their music library on shuffle, or to some playlist that Spotify or Apple Music cooked up? Is music really special when a computer algorithm is choosing what you’re listening to? Some argue that music experience is not as special in this case, while others think there is nothing wrong with algorithms.

these crazy but common things called algorithms

While they might seem like the most complicated things for those unfamiliar with computer science, or who are just plain bad at math (like myself), algorithms have actually been around quite some time – in theory. Modern formulations of algorithms did not begin to emerge until after the scientific revolution.

Algorithms can be generally defined as a “self-contained sequence of actions to be performed.” These actions can then perform various calculations, process data, and so forth. Or even more simply, an algorithm is just a set of instructions to be followed; some information, the input, is entered to be processed, and to be finalized as the output. Here’s an example of a simple algorithm:
Now, if you don’t understand any of what you just looked at, I totally understand. Either way, music algorithms are what are behind many of the playlists that we come across in music streaming apps like Apple Music or Spotify. They are the driving factors in these apps’ sorting programs, creating the plethora of mood and party playlists that we all know and love.

In fact, one of Spotify’s content editors, Austin Daboh, has said that, “we have 3 different types of playlists on Spotify…we have 100% handcrafted curated playlists…’algotorial playlists’…then we’ve got 100% fully algorithm based playlists.” Daboh claims that the reason why Spotify depends so much on algorithms is to account for the large multitude of users they have to cater to, and that it would simply be impossible to curate every playlist by hand. I suppose having this perspective makes sense when you have to administer such a massive streaming service like Pandora or Spotify.

There are even entire parties and raves based on people live coding on their computers, so the algorithms create the music people are dancing to. It’s called an “algorave,” and it’ s pretty cool.

Algorithms Aren’t everything, THOugh

Despite algorithms being a very practical thing within the distribution and the listening of music, I still don’t think that they are the ultimate future of music discovery and enjoyment. Sure, every now and then I like to log into Pandora just to see what they might throw at me, but the truth is, that despite making all types of “stations” that reflect a variety of genres, I still end up hearing the same songs over and over again. Pandora is run on algorithms, and while I admit it has improved significantly these past few years, they and other streaming services still disappoint me.

Human curated playlists always seem to be the most authentic form of mixing songs up perfectly, even if an algorithm can do almost the exact same job. Humans, with a natural intuition about what makes a good playlist good, will always beat out what an algorithm can create.