I imagine it’s only a matter of time that they could somehow develop photographic memory, not to remember what AI researches to create the images, but what AI creates itself.

Will it ever be able to retain memories, visuals, sounds to reuse and repurpose?

Notice how they never pan away from the shot, and if they do they never return? AI still can’t “remember” what was there before. Once the “camera” moves away, it’s gone. What’s there when it comes back could be something simple, but it will be different.

However, they still cannot exactly replicate the same person or character in a new scene. Every the microphone is different. You can tell they started with a prompt describing how the interviewer should look, but they change with every clip.

However, every video is still delivering this perfectly graded, contrasted shadows and colors, mystically and misty backgrounds that provide more of a mood than realism. The lips matching the vocals is getting better, but it’s not quite perfect. Hands definitely look better.