Reading about how it's done it's not so clear those other things are close at hand. Most of these SD things generate very low-res images and then use upscaling to make them high enough resolution that they don't look like crap.

Apparently the computational power required to make them larger grows exponentially (or geometrically?) so until we find new algorithms it might be a long time before the same technique can generate whole films and games and any coherent way.

Of course maybe that breakthrough will be announced tomorrow.