Cute! For Science!
Vocaloid, voice synthesis for music, is technically much easier for Japanese than English (I think I was reading there's like an order of magnitude difference between the number of discrete sounds or the number of discrete changes between sounds or something), so while English speech synthesis still sounds shit, and even japanese doesn't sound completely right, still they're making arguably (a lot of awful, awful pop, but also) some cool tunes. And it seems to be social media-type communities doing it.
I mean, for god's sake, they gave us
DANCEROID!!!!111!!!eleven. That Danceroid video is the archetype for hundreds of MMD copies using the software characters for the vocaloids... and their pantsu!!!
And
Danceroid r awesome. woooooo! \o/