Yes. These systems are working on a 3D problem in a 2D world. They have a hard time with situations involving occlusion. A newer generation of systems will probably deduce 3D models from 2D images, build up a model space of 3D models, generate 3D models, and then paint and animate them. That's how computer-generated animation is done today, which humans driving. Most of those steps have already been automated to some degree.
Early attempts to do animation by morphing sort of worked, and were prone to some of the same problems that 2D generative AI systems have. The intermediate frames between the starting and ending positions did not obey physical constraints.
This is a good problem to work on, because it leads to a more effective understanding of the real world.
How deep is the pool? Just because you have descended to depths that others were skeptical of doesn't mean there isn't a floor. Besides, these videos are less janky, but still obviously fake, for for sure they're cherry-picked for maximum effect.
Well - the videos in the link gave me same kind of response as literal Latin American gore videos, so line 3 to 5 still applies, and line 1 to 2 still applies to ChatGPT w/GPT-4 Turbo, so... I don't know what to make of this, maybe people like gore videos. Or something.
"I still haven't seen any text written by AI that doesn't contradict itself a few paragraphs later"
"I still haven't seen any picture made by AI that doesn't look like an abstract nightmare"
"I still haven't seen any picture made by AI that includes hands with the right number of fingers"
"I still haven't seen any videos made by AI that aren't janky and uncanny"
"I still haven't seen any videos made by AI that move me"
The transition from the first to the last took less than five years.