Slop.

What I hate the most about AI video generation is that it's very good at all the wrong things while being so random and uncontrollable as to be functionally useless. It can literally only make janky vapid slop and every improvement it gets only makes the problem more pronounced because it becomes more detailed and less glaring-error-prone without fixing the core problems that make it useless.

And the short in that post is like the ideal scenario for covering up those flaws, too: non-human characters with busy but uniform details no one can really pay too close attention to, that are always viewed from the front, and that remain in place while making overly exaggerated movements. And it still looks like ass with noticeable errors that literally cannot be corrected since the "creator" is just pulling a gacha lever over and over.