Music recommendation chatbot - yay or nay

Watch
Shoddy_Daikon
Badges: 2
Rep:
?
#1
Report Thread starter 6 months ago
#1
Do you guys like the idea of a chatbot (like Siri or Alexa) that can send you song/artist recommendations? hopefully it would help you discover some new stuff to listen to that's not just what Spotify/Youtube algorithms recommend.

Do you think it could also help out artists who don't get promoted as much on playlists on Spotify etc. ? I'm just kicking around some ideas.

Say you are in the mood for listening to something relaxing, you could ask the chatbot to send you some relaxing music. Or you could say you want to listen to a certain genre, then the chatbot could send you a song in the form of a Youtube video or spotify link.Thanks!
Last edited by Shoddy_Daikon; 6 months ago
0
reply
Arctic Kitten
Badges: 21
Rep:
?
#2
Report 6 months ago
#2
This is an interesting idea, but I have seen it poorly executed.
Being a person who likes really different music across the board, I often dislike what they recommended me. I like certain early 2000s punk music (certain song even, and don't like everything else from the artist), Sibelius (and find Beethoven mehhh), indie music, one or two pop songs, some Chinese movie song, and a few very mechanical sounding Vocaloid songs.
I believe this is what most people feel - they like a lot of different music, whatever sounds nice to them. A bot/ AI learning machine can only send 'what people like these songs listen next' but not identify your specific taste and adjust (at least to a good degree) unless you like generic pop songs like Taylor Swift and Ariana Grande.
I have been finding songs by putting random words into the search bar. It's actually working quite well. Mood-fitting musics.
3
reply
gjd800
Badges: 20
Rep:
?
#3
Report 6 months ago
#3
Nah, the last.fm algorithm does me fine for that. Or, yknow, asking my mates
2
reply
Shoddy_Daikon
Badges: 2
Rep:
?
#4
Report Thread starter 6 months ago
#4
(Original post by Arctic Kitten)
This is an interesting idea, but I have seen it poorly executed.
Being a person who likes really different music across the board, I often dislike what they recommended me. I like certain early 2000s punk music (certain song even, and don't like everything else from the artist), Sibelius (and find Beethoven mehhh), indie music, one or two pop songs, some Chinese movie song, and a few very mechanical sounding Vocaloid songs.
I believe this is what most people feel - they like a lot of different music, whatever sounds nice to them. A bot/ AI learning machine can only send 'what people like these songs listen next' but not identify your specific taste and adjust (at least to a good degree) unless you like generic pop songs like Taylor Swift and Ariana Grande.
I have been finding songs by putting random words into the search bar. It's actually working quite well. Mood-fitting musics.
Hm yeah I guess if you like particular bits of lots of different genres the algorithm would be confused... your method of using random words to find songs sounds interesting, maybe an AI could send you random songs too? Do you use words that express emotions to find songs?

What about something that tries to gender balance the artists you listen to? like if you always listen to all male bands it could send you similar bands that are female?
0
reply
username4313572
Badges: 17
Rep:
?
#5
Report 6 months ago
#5
(Original post by Shoddy_Daikon)
Do you guys like the idea of a chatbot (like Siri or Alexa) that can send you song/artist recommendations? hopefully it would help you discover some new stuff to listen to that's not just what Spotify/Youtube algorithms recommend.

Do you think it could also help out artists who don't get promoted as much on playlists on Spotify etc. ? I'm just kicking around some ideas.

Say you are in the mood for listening to something relaxing, you could ask the chatbot to send you some relaxing music. Or you could say you want to listen to a certain genre, then the chatbot could send you a song in the form of a Youtube video or spotify link.Thanks!
Sounds good 😊
1
reply
Shoddy_Daikon
Badges: 2
Rep:
?
#6
Report Thread starter 6 months ago
#6
(Original post by gjd800)
Nah, the last.fm algorithm does me fine for that. Or, yknow, asking my mates
What about something that tries to gender balance the artists you listen to? like if you always listen to all male bands it could send you similar bands that are female?
0
reply
gjd800
Badges: 20
Rep:
?
#7
Report 6 months ago
#7
(Original post by Shoddy_Daikon)
What about something that tries to gender balance the artists you listen to? like if you always listen to all male bands it could send you similar bands that are female?
Some people might be into that! I don't really care, tbh
0
reply
Arctic Kitten
Badges: 21
Rep:
?
#8
Report 6 months ago
#8
(Original post by Shoddy_Daikon)
Hm yeah I guess if you like particular bits of lots of different genres the algorithm would be confused... your method of using random words to find songs sounds interesting, maybe an AI could send you random songs too? Do you use words that express emotions to find songs?

What about something that tries to gender balance the artists you listen to? like if you always listen to all male bands it could send you similar bands that are female?
I tend to listen to band music with a mixture of male and female, I find it has better harmony and vocal range. So gender balance would not work well for me, but it's sound fun.
Though it's really difficult to find similar bands. Songs include melody, lyrics, and 'vibes'. Songs with the same melody can still create different vibes (looking at copyright music case). That's kind of the whole problem.
Youtube, and other music streaming company, work by identifying 'pattern', asking for a massive input, and giving tones of output for 'trial and error' kind of thing. Which means it doesn't identify deeply musical stuff, and thus unpopular music are hard to come by, even though fit perfectly with the taste.
0
reply
Shoddy_Daikon
Badges: 2
Rep:
?
#9
Report Thread starter 6 months ago
#9
(Original post by Serenity-M)
Sounds good 😊
how about something that tries to gender balance the artists you listen to? like if you always listen to all male bands it could send you similar bands that are female?
0
reply
Shoddy_Daikon
Badges: 2
Rep:
?
#10
Report Thread starter 6 months ago
#10
(Original post by gjd800)
Some people might be into that! I don't really care, tbh
do you think it might help to improve equality in music?
0
reply
Shoddy_Daikon
Badges: 2
Rep:
?
#11
Report Thread starter 6 months ago
#11
(Original post by Arctic Kitten)
I tend to listen to band music with a mixture of male and female, I find it has better harmony and vocal range. So gender balance would not work well for me, but it's sound fun.
Though it's really difficult to find similar bands. Songs include melody, lyrics, and 'vibes'. Songs with the same melody can still create different vibes (looking at copyright music case). That's kind of the whole problem.
Youtube, and other music streaming company, work by identifying 'pattern', asking for a massive input, and giving tones of output for 'trial and error' kind of thing. Which means it doesn't identify deeply musical stuff, and thus unpopular music are hard to come by, even though fit perfectly with the taste.
if it was used by people who don't have such a balanced taste, do you think it might help to improve equality in music?
0
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

Should there be a new university admissions system that ditches predicted grades?

No, I think predicted grades should still be used to make offers (549)
34.01%
Yes, I like the idea of applying to uni after I received my grades (PQA) (665)
41.2%
Yes, I like the idea of receiving offers only after I receive my grades (PQO) (327)
20.26%
I think there is a better option than the ones suggested (let us know in the thread!) (73)
4.52%

Watched Threads

View All