Why?
Iiuc this is just about APIs for the recommendation engine
It’s never been easier to generate recommendations (eg via LLMs and other routes)
The core functionality otherwise remains unchanged in the API
Iiuc this is just about APIs for the recommendation engine
It’s never been easier to generate recommendations (eg via LLMs and other routes)
The core functionality otherwise remains unchanged in the API
So happy I never got started on that little dream project that’s been knocking around the back of my head for a couple years…
LLMs require data, as I'm sure you know. This is locking up what was previously an interesting source of data, which undermines your argument over the long term
I honestly don't find Spotify's recommendations all that great. I definitely experienced a broadening (perhaps deepening) of my listening early on, but my experience has been that the recommendations are pretty shallow.
I find after throwing together a playlist with some stuff I like, it'll add a few more artists to my mental roster, then nothing. I'll get thrown around in the same loop with the same tunes and artists -- usually from the more famous albums.
I don't want to sound too much like the grouchy aging hipster that I am, but recommendations engines are just one of many ways of discovering music, and I feel like y'know, the old ways were better than just paying some company to do it for me. I'm talking here about being a regular on a local music scene, smoking weed with musicians, trading MP3s on the sneakernet.
Another thing where we just pay some money for "convenience", but are left with some hollow and empty algorithmic imitation of something we once loved.
Your LLM suggestion made me do a little sick in my mouth.
Maybe you have ondansetron around
Just what I want in a recommendation engine - to sift through garbage nonsense.
You know how people believe whatever they read, hear, and watch even though it might not be true? Well an LLM is something people read and to get over the hurdle of whether something might be true or good, you simply embrace it and ignore that it could ever be wrong. I don’t get it either as I get upset when I find out a source is mostly wrong.
As far as I can tell, the only thing actually any good at music recommendation is (some) humans.
I'm pretty sure it's not LLM based though, but rather domain specific, or possibly just a simple recommendation engine ("people who like x also often enjoy y").
This is in stark contrast to Spotify, where after a while it would invariably throw in some songs that just did not fit in at all with the rest, and it made for a jarring experience. I've never really had that happen with Apple Music.