Context
(wait for the tweet to load - it takes time sometimes)
The following embedded tweet (wait a little if its not loaded yet) contains a video with the caption "AI deepfakes will never be able to create this level of detail"
I will judge subjectively myself or with a group of people whether the current state of AI has reached a level where a video with a similar level of detail is possible to be created by AI before 2030.
This refers solely to the level of detail not to the video quality, which is quite poor in the video in the tweet.
The length of the video is 1:43 - the length of a video created by an AI system must be at least 60 seconds (1 minute) in length with a similar attention to detail level, or better.
Related questions
So this is a video that has been spliced together from independant episodes to convey some different meaning than original context. Would an AI that first does the generation of the independant episodes and only then does the splicing according to some prompt count? Or does it need to do everything in one go, without any splicing involved?
@ScipioFabius yes this specific video is just an example to give an idea of the level of detail. I'll probably pick and add more examples to make that clearer
Yes splicing counts