I think this will come, but it won't be competitive with filmmaking in a decade, not without real superhuman AGI. What's much more likely is efforts to reproduc

I think this will come, but it won't be competitive with filmmaking in a decade, not without real superhuman AGI. What's much more likely is efforts to reproduce performances will result in deep uncanny valley stuff - and correcting this last few percent for accuracy / weirdness will take a long time. Photorealistic video rendering is inevitable, voice duplication is already here (Tortoise should be much more well known but our culture is fixated on the visual side of this tech - https://colab.research.google.com/drive/1wVVqUPqwiDBUVeWWOUNglpGhU3hg_cbR https://colab.research.google.com/drive/1wVVqUPqwiDBUVeWWOUN...). But generatively creating a performance which accurately interprets the emotional context of a scene, in interrelationship with other characters doing the same? I don't see even a path towards that without AGI. At best you could crib specific performance elements and map them onto new models. My intuition is that to get all the way to generative movie you need AGI (or Zimbos so convincing we can't tell the difference). So what we will get in coming years / decades - the timeline is anyones guess - is movies acted out to camera in a rehearsal space, and that combined with a(n AI augmented) script to generatively create a film / character driven interactive game).