Cool & Wet

 

A friend suggested I do a how-to video on AI song generation. Isn't it all self-explanatory? Not really. I see it as a kind of "whispering". Sell You a Bridge took perhaps 25 "captures" which resulted in 12 curated songs, all the while refining the lyric. What's disappointing is that while the lyric might be finished in the final machine iteration, the music isn't. You're stuck with it. Certainly the way the words are "sung" isn't complete because a real person never interacted with them. A way around this is to re-generate the last iteration without the vocal and have someone sing it, but that changes it again. Every time you touch a piece of art or music it changes. Even listening to it with fresh ears changes it. A recording seems fixed as a painting is, but the internet makes the paint wet again. The McLuhan corollary is that the digital art made on the internet and with the internet makes it "cool". In music, the "hot" version was always the score or sheet music, but the interpretations made it cool.