On: Imagen, a text-to-image diffusion model

[exac]

You can sort of do that with https://fairuseify.ml

[jrh206 → exac]

I believe that this tech is possible, but this site doesn't provide it. Look at the source of the page: it's just a bunch of sleeps and then you 'download' the same file you provided.

[waqf → jrh206]

The tech may be possible, but it won't solve anyone's copyright problems. The result would be a "derived work" of the original, irrespective of whether it sounded similar or not.

[aembleton → exac]

I tried that site and the music sounds the same. I wonder if you can use this to bypass YouTube content ID check.