Model: Veo 3 Fast
1280x720
24fps
00:09:27
From a film shot in a single take, a single image is collected every eight seconds - the longest length that the video model used could generate. The sampled images are used as references for the start and end of these eight second chunks, and the video model invents the motion that connects the two. Every eight seconds, indexicality is restored for a single instant before deviating back at a constant, sinusoidal rhythm into the flow of artificial imagination.