A Bot Made Me a Mixtape

Howdy folks, and welcome to a new section of my blog, Soundtrack. This will be completely in English and it will deal just around music topics.
Mostly you'll be finding album suggestions, review, new releases plus some deep thoughts on streaming services which I'm very fond of (Pandora, Spotify, Rdio, Tidal... you name it, I've tried it). 
Finally, I'm Italian, I've studied English for several years now, but bear with me in case you'll find written mistakes that could bother you a lot. I will try not to make any. 

Ok, so, without further ado let me dig into the main topic of this first post: Discover Weekly by Spotify. 

Discover Weekly has 1 year life now and for those of you who still don't know what it is, it's a super killer feature that creates a personalized playlist based on your music tastes, updated every Monday.
It's so perfectly crafted that already 40 million users have already used it and, as said, it is the current best differentiator of the music streaming service. 

But it is still crafted by a computer. Better by Spotify machines or more precisely by the recently acquired company Echo Nest (The Ringer has a good article on Echo Nest) which system was once one of the strong points of the beloved and a-lot-missing Rdio service that decided then to ditch it after it has been acquired by Spotify. 

So far, so good. The system is fine, delivers what it promises and normally allows me to discover at least 1 new artist at every other week pace. Furthermore, is the only reason for me to maintain a Premium account on Spotify, because it's the only way to listen to Discover Weekly on my Sonos system. A part from that, after the farewell of Rdio, I found my perfect spot on Apple Music (but that's another story). 

My title hints at "I Made You a Mixtape" of  Federico Viticci on MacStories. He tells the story on why Discover Weekly is becoming more and more important to him because: 

Spotify's Discover Weekly, on the other hand, has brought back the edge I was missing from music streaming services. Discover Weekly is helping me find new music like I haven't done in years. Every Monday, I feel like I'm back in high school and staring at the handwritten back of a mixtape cover. I crave its sense of discovery.

Fair enough. Music discovery, like music taste, has no rules, it's something very personal and most of the times does not follow any specific path.
The way I discover new music (an activity I'm really addicted to) is a multifaceted process. It goes from Discover Weekly, as said, to the For You section on Apple Music, the well curated A-List playlists on Apple Music based on different genres, reading blogs or specialized websites & magazines and so forth. 

But the past Monday one of these pillars was broken down by something I didn't expect would have worked this way.

Last week was bank holiday in Italy, so I decided to take couple of days off for the celebration of the Italian Republic Day. I had some driving hours ahead of me, so I've to save on my Spotify library some suggested playlists with some party music was kind a ritual and a must have to me. 
I must start by saying that this is defenetely not my genre of music. I'm addicted to Hard/Alt/Indie Rock (Just take a look at my 3 collections) but having some other people on board, it's always fun to have that late '90 early '00 kind of mood where we all know the songs that are playing and sing on those notes. 

Listening to a set of songs casually, very far from my music tastes, I personally think doesn't mean that I like those. It doesn't mean that I'm starting to listen to those regularly, because I didn't take any actions against them, such as: hit a button to like every single song or save into my personal library. 
I've just subscribed to a suggested playlist and saved it offline not to waste 3G data connection while abroad. That's it.

But here what happened this Monday when I opened my Discover Weekly to listen to the list of songs as I used to do for 1 year now. Look at this list below, click on it to zoom in and better read what kind of suggestions I got...

My altered Discover Weekly of this Monday

That is not me, this is a playlist of a dancefloor addicted or a person particularly attached to a late '90 melancholy.

A Bot Made Me a Mixtape

And it failed, how possibly 2 hrs of listening to a particular music genre could have screwed up 1 year of listening and usage of Discover Weekly?
This sounds totally a bullshit to me. Really. I can understand that sometimes the songs filling my "personally crafted" playlist won't encounter my attention and tastes, but this is a mistake by design. 

Apple Music, on the contrary, may not have yet a killer feature like this one, but all its "For You" sessions and the so well built suggested/personalized playlists are just one โ™ฅ away from you. And, sorry to say, it is so much more reliable. 

A screen from Apple Music UI on iOS

A screen from Apple Music UI on iOS

Now, I don't know if my Discover Weekly will be fixed by the morning of the next week's Monday (my good friend Lorenzo says yes), but crucial fact remains, a machine is entitle to build its supposed best differentiator feature just on what you strictly listen and processes that as it as your preferred genre over all. 

So imagine you have kids, you want to hold a themed party or whatever situation you might think of where music is involved, and use your very personal Spotify account to reproduce it, you are basically screwed up because you know...next Monday your Discover Weekly is going to be populated by weird stuff.

So for Spotify, and how its algorithm is working, listen to music should be considered as an active action that states what you are and stand for no matter what are listening to, while I still remain convinced that it should be decided by the end users what they like or not, and the so-called machine learning should start processing from that set of data and propose a coherent offering of tailor made playlists. 

It's a long path the one that will bring machines to learn us and in this case is demonstrated that even the most sophisticated music software yet isn't able to interpret what a person really wants at the end of the day. 

Although, future is bright and we cannot expect nothing but better and better experiences.