Exploring video generation non-linearity with SVD

We took the opportunity to learn to use ComfyUI and Stable Video Diffusion. We were able to install ComfyUI environment on our own hardware and made it available for the whole team. Starting from the basic workflow and generated a short clip and 4 different alternative clips. So we are exploring kind of exploring "parallel realities" from the original image. This idea originated after listening to a Karen Palacio's videoessay in which she mentions the parallelism between non-linearity in dreams and machine learning generative systems. https://www.youtube.com/watch?v=gA_lfIddqiA We also had the idea to explore different narratives in the "parallel realities" based on Kurt Vonegut's lectue "The Shape of Stories" and variations on that theme.https://www.youtube.com/watch?v=oP3c1h8v2ZQ Unfortunatelly Stable Video Diffusion is not easily conditioned with text yet. And we didn't have enough time to explore this idea.