The research team at the National University of Singapore has successfully developed a free AI video creation tool called Show-1. This tool employs a hybrid architecture of pixels and latent models to generate high-quality videos from text descriptions. Compared to other methods, Show-1 only requires 20-25% of GPU memory but excels in realism and text alignment. Additionally, Show-1 offers its open-source code and model weights for free use by the research community and developers.