Found this interesting.
Interesting indeed.
One thing that caught my attention about AI in music was Bjorn’s comment about AI being trained on existing music.
Aren’t we all trained on existing music?
You see a lot of interviews with musicians where they talk about other artists who inspired them. Wouldn’t that be considered a form of training? And if so, based on what Bjorn said, shouldn’t there be a means of compensating the artists who were the source of the inspiration?
I hope the answer is no, otherwise I owe Jimi Hendrix and BB King a butt load of money for ripping off a lot of their licks. 
The speech to text algos aren’t quite there yet [see text at bottom of image]
.
.
Lol wth?
Oh, “…A minor with Aegean Bays” = A minor with a G in bass
“F see with a Nina base” = F, C with an E in the bass