Aiva is way better than this honestly, and I mean the pure AI part of it, that (AIUI) only really generates monotimbral, piano-like music (the orchestration in

Aiva is way better than this honestly, and I mean the pure AI part of it, that (AIUI) only really generates monotimbral, piano-like music (the orchestration in the full pieces they release was - by their own admission - done by humans, last I looked into it). You can actually tell that Aiva involves real serendipitous ("out-of-sample") creativity of the sort that AIs (and good human composers!) are best at. (But the openly-available models I mentioned elsewhere in the thread are still a bit better, TBH. At least IMHO, and when it comes to rewarding active listening.)