21–30 of 85 posts
I've been a big fan of AI music for background noise while coding
AI music rocks. Especially the ones made on Udio, which I think it's the best model as of right now. I hope we soon have a "Flux" like equivalent model, but for music, that you can run locally/fine-tune,etc...
In the end, made by inspiration or by some algorithm, following patterns, marketing studies or whatever, by AI or not (if there are cases you can tell), what matters is your experience.
Why "awful"? In what is different AI generated music and organic, free range, supposedly human made one? I mean, how do you tell what percent of today's commercial music has been generated more with a calculator than through inspiration or whatever? In the end, made by inspiration or by some algorithm, following patterns, marketing studies or whatever, by AI or not (if there are cases you can tell), what matters is y…
It would be cool if the author shared the AI music channels list. I love AI music and would greatly appreciate some recommendations of channels to listen to.
In ten years this will be as quaint as a blocklist for music that uses samplers would have been in 1980. Not to say there’s no value today; most pure AI music is terrible and I don’t want to listen to it. But that’s a very temporary state of affairs.
It’s ridiculous that YouTube doesn’t allow blocking of channels. I find the prevalence of fake movie trailers to be absolutely infuriating, and I do not understand why YouTube and the IP holders don’t nuke those channels, or at least let people block them!
The only way I can stomach YT at this point is with blocktube and unhook; I can disable parts of the interface, block channels and content individually or with regular expressions, and filter comments with regex - a necessity given how many scams are being run right now.
I dread the day they drop RSS feeds for channel updates.
https://github.com/amitbl/blocktube
Edit: also see how YT search gives you unrelated nonsense. User-hostile may as well be their mission statement at this point.
In ten years this will be as quaint as a blocklist for music that uses samplers would have been in 1980. Not to say there’s no value today; most pure AI music is terrible and I don’t want to listen to it. But that’s a very temporary state of affairs.
It's more like a blocklist for the demo programs from samplers.
Earlier quoted context omitted.
Same, though by surprise. I've been generating a very specific uncommon genre of music on Suno (a sort of Slavic accordion drum'n'bass - a totally random idea I had one day) and found myself listening to it more than I anticipated. I dislike most of Suno's vocal output but the instrumentals can be quite good especially in unusual fusions of genres.
> sort of Slavic accordion drum'n'bass Interesting. Do you have an example?
I quite like music with uncommon modes and weird scales and Suno isn't too bad at emulating this - especially with Arabic music (e.g. "swirling dramatic orchestral arabic music with oud guitar" will yield many good, if stereotypical, results). I'm not familiar enough with traditional regional music to pick out all the flaws so it works for me, but a "local" would probably say it sounds unrealistic in the same way that Suno-generated typical Western pop and rock sounds off to my ears.
If you manage to curate a large enough collection of AI-generated playlists, would you consider putting it on Kaggle? This way, others could use it to train AI models to automatically detect AI-generated music. We did something similar by sharing a dataset of around 7,000 AI-generated fake podcasts on Kaggle: https://www.kaggle.com/datasets/listennotes/ai-generated-fak...