On February 15, 2024 OpenAI announced the launch of Sora, a new milestone in the development of video-generative AI technology that produces smooth and realistic videos up to one minute in length. According to the company's report, Sora uses Spacetime Patches and Diffusion Transformer as its core technology, utilizing the generative power of Diffusion and the self-attention mechanism of Transformers to generate video content by predicting clean blocks, and then processing the spatial and temporal patches with Transformers. Transformers process spatial and temporal blocks.