That's because the recommendation engine that Last.fm used back in the day was made the incredibly expensive way: the entire corpus was hand-tagged and cross-linked by humans atop an enormous CDDB. Last.fm, Audioscrobbler, and MusicBrainz (the association engine) were all linked together.

I switched to Apple Music to save some money and I find the curation and the recommendations to be significantly better than Spotify.

Cannot call lastfm algorithm advanced in any sense. Just opened Amon Tobin page: "similar artists: Kid Koala and DJ Kush", which is an impressively shallow understanding of the last 20 (!!) years of his life, and this happened with almost every artist on the platform, because the average sum of tastes of every listener does not exist in reality. E.g. in the case of Amon Tobin, Kid Koala is the average of similarities between early albums and recent releases, which is just not true, his music cannot be averaged throughout his career. I love my Web 2.0 youth, but the average similarity algorithm doesnt deserve praise. Its not better, its nostalgia and lack of faang-style unlimited greed which confused with better quality

Edit: of course spotify-style recommendations are much much worse, I just mean that lastfm doesnt have good algorithm either because artists are not consistent in releases. What is an average between electronic cult classic "The last resort" and every other Trentemoller album in strict indie rock style? This average does not exist

I'm 90% sure that music labels pay to "put their thumbs on the scales" with these recommendation algorithms in order to push their "hot" artists. I wonder how many of these problems are a result of that.

The other frustration I’ve noticed is that they key in very heavily on artist and specific “genre” designation as what feeds the recommendation, which is actually quite bad for anyone who likes experimental work.

I understand that if your recommendations are based on “people who like this also tend to like that” then you’re right in the strike zone. But that approach is basically agnostic to any property of the music itself. Suppose there’s a rock band that released a specific song where they’re experimenting with a new style that has an atypically (for them) funky/jazzy influence. If I say I want more songs like that I mean songs that fuse rock/jazz/funk, not more songs that fans of [rock band] are into.

I still think for new music discovery Pandora’s approach remains the best if you really curate a station for yourself. Apple Music has been good for creating very listenable playlists though, and their new AI playlist generator has been very fun. Surprisingly, YouTube also seems to have some secret sauce where they recommend a lot of interesting stuff that I’ve genuinely never encountered before. I suspect this is because there’s a lot more amateur and experimental artists on there doing weirder stuff and it’s able to find audiences for those in ways that the music-focused services have less visibility into since their catalog is so focused on stuff from the recording industry.

I just remembered that I met one of my best friends to this day through Last.fm. It was 2009 or so, and you could leave messages on concert pages.

I posted asking if anyone wanted to go with me since I didn't want to go alone, and she sent me a message.

Good times.

I second this. I started as an Audioscrobbler user before the Last.fm merger. I have tracked nearly every track I've listened to for 21 years. It's awesome seeing how my habits have changed over the years.

As a long-time user, I do enjoy seeing how my tastes have changed over the years and which artists and albums I play the most. I also tend to agree that the Last.fm recommendation engine was perfectly fine for my use case compared to the algo that Spotify uses now. https://www.last.fm/user/wyclif

Same. I have one or more gaps in there which I wish I could go back and correct. I feel like integration with the service is a must for any music thing I pick up, the most recent being this year, resurrecting my old iTunes library via Navidrome.

More than a feeling.

Pretty much all the machine learning recommendation engines that emerged in the Netflix era were doomed to collapse under their own weight over time for non-mainstream users because the some limited number of mainstream modes dominate as most statistically "optimal" across the total user pool. These algorithms are best in the early days, when they're still exploring the content space for good novel fits but eventually get trapped into deep, boring grooves that work really well for tons of non-discriminating users with similar tastes.

Separately, in real commercial terms, they're all fundamentally poisoned by business model objectives of highlighting cheap content or servicing partnership/advertising deals, etc. And that problem also becomes more and more prominent as the companies running them grow and become more influential and as they need to squeeze harder and harder for revenue growth.

It was basically just a long, winding, wildly expensive road back to broadcast radio programming.

It was a good run for a while, but we're long due for a new model.

Yep, member since May 21 2005 here, still scrobbling with Spotify. Don't think I've ever used any of the radio features on the site, really; even back in the 00s all I used were the WinAmp/Foobar plug-ins.

If you're a Spotify user, you can get even more precise data by downloading your listening data. The website I linked gets data from MusicBrainz and tries to fill in the gaps with an average, but even then it gets some things wrong.

E.g. Fishmans - Long Season is a 40 minute song, but the website's considers it as divided into 4-5 parts. And you don't have to listen to the full song to get a scrobble.

In the Spotify data you get the exact number of seconds you listened to it. And it is surprisingly complete and easy to use too. With LLMs I bet you can load it into pandas and construct queries for any insight you want in seconds.

You can install your spotify and pull in all the data from Spotify.

https://github.com/Yooooomi/your_spotify

Originally, it kind of worked like radio; it curated music for you, you could like, comment or skip tracks. It'd reinforce the algorithm, and you'd start finding great artists. I liked the Blues catalogue a lot, even though I was listening to reggae, ska, punk, etc. It just seemed they had the best music catalogue. I remember checking how big the catalogue was, comparatively with others, which was much smaller, but much, much better!

Today, we have Generative AI, generating an incomprehensible number of songs that no one will ever listen to.

I don't remember if I had to pay for Last.fm or not back then, but I'd definitely pay to have access to that old system.