yeah but how can we train that. even if we show it the best anime artists that use certain technics. it can copy style and it may can follow up on animation. but real high end animations im thinking of are set together in your brain. if you look at them on how they are animated frame by frame they make no sense. for the coming years ai will do amazing things. but i think from the current transformer based, diffusion and reconstructing we may just get another style filter that looks cool but not 100% fitting. i think we may not be able to shake that unnatural feeling that comes with those ais for quite some time. but happy to be wrong