The video didn't really deliver on its title (a riff on the DESTROYED youtube clickbait meme.) He seemed to be arguing that "1) Pop songs are too robotic soundi

The video didn't really deliver on its title (a riff on the DESTROYED youtube clickbait meme.) He seemed to be arguing that "1) Pop songs are too robotic sounding, 2) listeners don't care, 3) that will make it easier for AI to take over, and 4) I don't know how musicians will get paid for AI-generated songs." (The last point recalling historical worries that player pianos, recorded music, radio, drum machines, synthesizers, or other music technology would displace musicians.) He's almost certainly right that AI will be used in music production, but he didn't seem to present much evidence for point (3) beyond "I can certainly imagine AI making a song like this." And what is to prevent AI from being trained to sound more organic or "human"? Drum machines already include swing and randomization (or "humanization") features, and adding varied swing and 2 vs. 3 vs. 4 to a metronomic beat was popularized decades ago. Regarding autotune itself, Cher (in 1998 as he notes - this isn't new!) and T-Pain innovated by popularizing it as a vocal distortion effect beyond simple pitch correction. (And as Rick noted T-Pain is actually a good singer and also licensed the fun I am T-Pain iOS app to recreate his signature effect.) At the end of the day autotune and time quantization are production tools which can be overused or misused like anything else. They can certainly make vocals sound unnatural, bland, mechanical, or distorted. But robotic voices and/or mechanical beats have been heavily used in popular music for decades in the case of vocoders, click tracks, and drum machines - or centuries in the case of the venerable metronome. Compare his defense of backing tracks: https://www.youtube.com/watch?v=SZXUPuiC9Eo https://www.youtube.com/watch?v=SZXUPuiC9Eo