Live data from Hacker News

Last.fm turns 20

theverge.com

51–60 of 178 posts

Re: Last.fm turns 20

#51

Still use last.fm. Most "You might like this" algorithms go straight for the low-hanging fruit, and often fail to take any kind of nuance into account. I can't tell you how many of the streaming services will see my Black Sabbath play history and immediately recommend, "If you like Black Sabbath, you should love...Slipknot!" But I've never had a real person make that mistake, because a real person who looks at my las…

I had always assumed that recommendation algorithms didn't work on genres, but instead did a lookup based on something like:

"You like X" "Person Y also likes X" "Person Y also likes Z" "I should recommended Z to you" (proportional to the number of people there are who liked X and Z, compared to other options)

If this is the case (and perhaps I'm wrong), it's nothing to do with the algorithm not understanding the subtleties of the genre, but it's actually just that on average those subtleties disappear because they are unique to you. (Forgive me, I don't know your genres at all, but), I would guess that on average people who like Black Sabbath really do typically like Slipknot as well, and that is why the algorithm recommends them to you.

It's not that the algo is any worse than human recommendations, it's just that it's an average of all humans likes, which will never fit your exact personal unique preferences because they aren't you.

Re: Last.fm turns 20

#52

Still use last.fm. Most "You might like this" algorithms go straight for the low-hanging fruit, and often fail to take any kind of nuance into account. I can't tell you how many of the streaming services will see my Black Sabbath play history and immediately recommend, "If you like Black Sabbath, you should love...Slipknot!" But I've never had a real person make that mistake, because a real person who looks at my las…

Yeah, music and other such automated recommendations are truly infuriating - I like to imagine it is an attempt to tweak people instead of helping them. This belief is reinforced when often these recommendations don't come with a 'never recommend again' or 'stop recommending altogether, please' option. I am grateful though that such 'helpful' behavior is often successful in helping me get off the service in question and not use it again.

Re: Last.fm turns 20

#53
post #8

Last.fm was cool and had huge community but cbs bought it and rewrote it making it slower, clunkier and removing features, essentially destroying it. Back in the day practically all players (including spotify) had scrobbling support.

I had even managed scrobbling from the iPod classic. There used to be a windows app that would keep track of all the music listened to since and would scrobble the bulk, in batches.

Re: Last.fm turns 20

#54
Last.fm died (as in for me personally) when I stopped cultivating my own music library. I used to have gigabytes of MP3s and FLACs, all neatly organised into folders, usually by artist/album, and meticulously maintained ID3 tags. All played through software like Winamp with the audioscrobbler plugin, or iTunes when I had my beloved iPod classic.

All of that drifted away as I got older, and the dawn of streaming services like Spotify came onto the scene. I'm not sure where my music is now, probably on a hard drive somewhere, dumped amongst other junk.

I think spotify used to come with last.fm support but I think I just lost interest in the whole thing, I don't consume music in the same way as when I was a younger man.

EDIT: Just logged into my last.fm account and it looks like the scrobbling still works from spotify, so it's been scrobbling all this time, probably for 10+ years without me logging in!

Re: Last.fm turns 20

#55

Still use last.fm. Most "You might like this" algorithms go straight for the low-hanging fruit, and often fail to take any kind of nuance into account. I can't tell you how many of the streaming services will see my Black Sabbath play history and immediately recommend, "If you like Black Sabbath, you should love...Slipknot!" But I've never had a real person make that mistake, because a real person who looks at my las…

> Human review and recommendation still beats algorithmic recommendation by a mile if you have discerning tastes

I completely agree. But last.fm doesn't do human review either, right? OTOH, I think the "radio" feature on spotify is wildly underrated. I used to spend hours compiling playlists for different moments/social settings/moods (still do, because it's fun), but picking any song/album/artist and then getting an endless stream of similar music is an amazing feature. Unfortunately, sometimes it screws up and will suggest me the same songs over and over again and there is no way to get a little more control over the algorithm.

I came to hate the Superstar cover by Sonic Youth and "I'm So Tired" by Fugazi for those reasons. Great songs, but they just kept coming up as the first recommended track and I can't stand hearing those intro piano chords on the latter anymore!! According to my last.fm [0] I played it 50+ times and I don't even remember once seeking out that song or putting it in one of my playlists.

https://www.last.fm/user/mahathu

Re: Last.fm turns 20

#56
I remember being at University with one of the founders (maybe late founder?).

I recall him working on audioscrobbler as a 3rd year project, and was quite impressed with it, mostly by the fact that it actually worked and recommended me something I liked.

It maybe started as a plugin for one of the popular audio players at the time (name completely escapes me though).

Only other thing I remember was him being super passionate about it, and wanting to carry on development after uni.

Re: Last.fm turns 20

#57

Still use last.fm. Most "You might like this" algorithms go straight for the low-hanging fruit, and often fail to take any kind of nuance into account. I can't tell you how many of the streaming services will see my Black Sabbath play history and immediately recommend, "If you like Black Sabbath, you should love...Slipknot!" But I've never had a real person make that mistake, because a real person who looks at my las…

I had always assumed that recommendation algorithms didn't work on genres, but instead did a lookup based on something like: "You like X" "Person Y also likes X" "Person Y also likes Z" "I should recommended Z to you" (proportional to the number of people there are who liked X and Z, compared to other options) If this is the case (and perhaps I'm wrong), it's nothing to do with the algorithm not understanding the sub…

The term for this is https://en.wikipedia.org/wiki/Collaborative_filtering

Re: Last.fm turns 20

#58
All these music services today suck compared to last.fm.

I’ve got like 30k songs played and discovered on last.fm and I stopped using it when they ended the service as it used to be years ago.

Re: Last.fm turns 20

#59
post #45

Earlier quoted context omitted.

Me too! 480,000 scrobbles and still going. I recently downloaded all of my scrobbles so that at least I'd have the listening data in case the service ever goes down.

I did the same recently after seeing it was possible. 150,000 scrobbles for me, which I thought was a lot until I saw your count! https://www.last.fm/user/carldr

I've seen users with 1.7 million scrobbles on Libre.fm (which means they imported their scrobble history)

Re: Last.fm turns 20

#60

Still use last.fm. Most "You might like this" algorithms go straight for the low-hanging fruit, and often fail to take any kind of nuance into account. I can't tell you how many of the streaming services will see my Black Sabbath play history and immediately recommend, "If you like Black Sabbath, you should love...Slipknot!" But I've never had a real person make that mistake, because a real person who looks at my las…

I had always assumed that recommendation algorithms didn't work on genres, but instead did a lookup based on something like: "You like X" "Person Y also likes X" "Person Y also likes Z" "I should recommended Z to you" (proportional to the number of people there are who liked X and Z, compared to other options) If this is the case (and perhaps I'm wrong), it's nothing to do with the algorithm not understanding the sub…

> It's not that the algo is any worse than human recommendations, it's just that it's an average of all humans likes, which will never fit your exact personal unique preferences because they aren't you.

But that is a difference though, because humans possess empathy: algorithms don't. Step away from music for a second and consider movies: if all your algorithm does is recommend movies because Person X likes Movies 1, 2, and 3, it will take many layers of metadata tagging before the algorithm figures out that it shouldn't recommend Movie 2 to Person Y (even if Movie 2 is a star-studded, Oscar-winning classic), while a human being could probably spend 30 seconds looking at Person Y's viewing history and realize, "Ah, this person has very few action movies and zero horror movies in their viewing history; they probably don't like violence. I would not recommend they watch Movie 2."

Post reply on HN