I had an idea to find a multimodal or model that can listen but not just for SST/transcription, but to sort of describe/understand/interpret what it’s hearing—- and then have it listen to my entire music library, and then have it peg my exact music taste and recommend other music, or to just do something interesting or fun with what it has learned or interprets my library. I don’t think there are any multi modal models that could do this though?
Edit, just found a Reddit thread from 9 mos ago about this concept. Interesting, wonder what the current state of this idea is. Also, Pandora, Google/YT music have been able to do this for years— eg recommend music based off user taste/similar songs etc… I wonder how they are doing it? Obviously some sort of AI/algorithmic thing I would think, but not LLM related I wouldn’t think.
LLMs don't "have taste" - they statistically model human preferences from training data, mapping features to popularity patterns without experiencing the music itself.
Indeed. I'm deeply skeptical of the proposition that LLMs can have taste at all. However, there's an even more fundamental question: is "taste" an objective thing in the first place? If not, then it has no meaning to say anyone's (or anything's) taste is "good" or "bad". The most you can say is how similar another's taste is to yours.
Saying a non-multimodal LLM (like many of the ones in the article) can have "taste" in music is like asking a blind person to critique pieces at the Louvre based on their descriptions.
Exactly, I feel like at the very least, you need to have the model listen to the music/song and have it see if it “likes” it etc. I don’t think the OP did that unless I missed it? Also see my other comment.
I had an idea to find a multimodal or model that can listen but not just for SST/transcription, but to sort of describe/understand/interpret what it’s hearing—- and then have it listen to my entire music library, and then have it peg my exact music taste and recommend other music, or to just do something interesting or fun with what it has learned or interprets my library. I don’t think there are any multi modal models that could do this though?
Edit, just found a Reddit thread from 9 mos ago about this concept. Interesting, wonder what the current state of this idea is. Also, Pandora, Google/YT music have been able to do this for years— eg recommend music based off user taste/similar songs etc… I wonder how they are doing it? Obviously some sort of AI/algorithmic thing I would think, but not LLM related I wouldn’t think.
https://www.reddit.com/r/udiomusic/comments/1gzvk70/a_llm_th...
“Can LLMs have taste in music?” would need to be answered first, esp if you want to argue they can.
LLMs don't "have taste" - they statistically model human preferences from training data, mapping features to popularity patterns without experiencing the music itself.
Indeed. I'm deeply skeptical of the proposition that LLMs can have taste at all. However, there's an even more fundamental question: is "taste" an objective thing in the first place? If not, then it has no meaning to say anyone's (or anything's) taste is "good" or "bad". The most you can say is how similar another's taste is to yours.
Saying a non-multimodal LLM (like many of the ones in the article) can have "taste" in music is like asking a blind person to critique pieces at the Louvre based on their descriptions.
Exactly, I feel like at the very least, you need to have the model listen to the music/song and have it see if it “likes” it etc. I don’t think the OP did that unless I missed it? Also see my other comment.
As they say, LLMs don't need to have good taste in music, they just need to have taste in music as good as typical humans.
With that, we no longer need humans to listen to music!