I gave it a shot… and got an interesting failure.
I chose Cyndi Lauper’s “Time After Time” mixed with Foxy Shazam’s “Oh Lord,” because they’re songs I love and well, why not?
The AI’s transitions are clumsy, and it doesn’t seem to be able to strip the vocal line from its background. If you listen closely you can hear both song’s lyrics at the same time, one loud, one soft. Since one song is shorter than the other, when it ends, the other song is abruptly cut off.
And yet… there’s something there that’s sooo close to right. With a human’s touch, this could be a successful mashup. If I had the slightest idea how to do it, I’d try it myself, but I’ve got no clue. 
