thinking things, not anything sensible enough to call a post, but stuff i might want to remember. call it memos to myself.
one – being manicesque last night when i shoulda been sleeping, i was thinking again about how the raddest music is all epic bombastic, but not sustainable for a full album. if i ever tried to make music, it might be fun to take those over the top songs from different genres and try to match their level of ruckus, make a multi-genre showcase of over-the-top musique. could digital underground’s doowutchyalike, bonnie tyler’s i need a hero, and judas priest’s sentinel ever share an album?
two – find out if fables is anything i’d ever want to do anything with, bc hey, free IP lol.
–
This just in: Stability AI has released a model that generates sounds and music from text. So in principle you could ask for “a zombie chorus fronted by adrian belew with john bonham on drums” but it’s not that good.
The question in my mind is whether it’ll be good enough to completely crater “pop music” with a flood of AIpop. That would be fun! Release the kraken!
Here’s one – sit back and listen to a music auto-generator, never the same song twice, 27 hour long jams… Who knows what the future will bring?
The possibility of getting decent quality generic art from machines, from music to images, should emphasize for people that part of what we like about art is the persona of the artist coming through. Not that you couldn’t make artificial personas, but it’s cool to get a piece of a real person’s soul, such as that concept be. I hope society isn’t too jackassed to throw out the cool useful tech in misguided attempt to protect the special pooky artists that they failed to protect in any other way for over a crapitalist century now.
GAS @ # 2: … a music auto-generator, never the same song twice, 27 hour long jams…
William Gibson posited exactly that, dub reggae style, in orbit of course, in Neuromancer, 39 years ago.
dub reggae lends itself to that idea very well, heh. good call, billy.
–