I think it's a lot better at classical, orchestral, and instrumental music than it is at anything requiring vocalization. I created this in less than 20 minutes: https://app.suno.ai/song/eb93c25b-bdbe-4c9f-8e03-66e9479c869...

I need to stem it, fix it up a bit, and remix for stereo in a DAW but it's much better than I expected for my first ever piece of music. Obviously it'd take a lot of work to create a Hans Zimmer level OST from the tool but IMO it wouldn't feel out of place on a Ludovico Einaudi album or on some Spotify or Pandora classical radio.

That's actually a very good piece. Like something I'd hear on late night Paradise Radio. If I was creating an indie movie on no budget I'd be all over this technology for the soundtrack.

I don't think musicians and composers are going to disappear as a consequence of this technology, in the same way that theatre actors were not made obsolete by film. What I do think is that a whole new category of professionals will be created - musicians and composers who get paid to train AI models. I bet it will pay better then the laughable amounts that are streaming royalties.

At some point in the future, wanting no part in AI-generated content is going to be like that old Onion headline. "Area Man Constantly Mentioning That He Doesn't Even Own A Television".