Originally Posted by Fdemetrio
So we can simulate Sinatra, all we need to do is put on one of his albums,. and we already have him. No Ai needed. We can even isolate his vocals, and chop up his lines so he's saying new things without him ever having said them. And it's HIM saying it.

But the real point is why are we simulating Sinatra? Cause he's Sinatra. An AI superstar, nobody will care about.

The main thing I see Ai doing is allowing anybody, musical or not, to create a song and use it for their podcast or, their jingles or for even in movies. Background music could be changed drastically.

It's not going to change the one on one exchange between human beings. I play, you listen, you understand me, you like it. You want to see me peform because you see yourself up on stage. THAT is not going anywhere.

But it will probably affect and limit the ways creators of music can make money doing it.

Dom,

Just so we're clear, it's not just Sinatra but several thousands of musical artists going back to Enrico Caruso(#69), to Dylan(#608) and Springsteen (#844), and into modern acts. Anyone you can think of.

Here is one list of about 4000 recording artists:

https://github.com/openai/jukebox/blob/master/jukebox/data/ids/v2_artist_ids.txt

There are a million uses for something like this. I personally wrote several songs throughout my life with real recording artists in mind. Now I can have a faux version of them actually singing those songs. C'mon, Dom...you've never written a song with Springsteen in mind as the singer?

I have created song "gifts" that sing directly to certain friends using AI versions of their most beloved recording artist's vocals, and in the sung lyrics calling my friends by their names, saying personal things about them only they would know...

I am currently creating an "opera" using dozens of different OpenAI Jukebox singers. It is the most enjoyable work I have ever done. It uses an actual movie in the public domain, and I am swapping out all the original audio with audio I am creating, partially with OpenAI Jukebox, lip syncing the vocals I am creating to where there was spoken dialogue originally.

Podcasts and jingles and background music, yes. But consider that maybe, if you chose to go down the rabbithole and look more closely, you might find there's a far greater potential for such an AI driven software...

Mike

Last edited by Michael Zaneski; 02/22/23 05:57 PM.
Fate doesn't hang on a wrong or right choice
Fortune depends on the tone of your voice

-The Divine Comedy (Neil Hannon)
from the song "Songs of Love"
from the album "Casanova" (1996)