14 hours Ago By Jesus Rodriguez
Video is rapidly becoming one of the next frontiers in generative AI. From OpenAI’s Sora startups like Runway or Pika to the new efforts from Google, most of the large video generation models remained closed-source. There are open-source efforts such as Stability AI’s Stable Video, but the quality is not at the same level as that of the large commercial alternatives.
The barrier of entry for pretraining video models is fundamentally larger than language or image, as datasets are more scarce and difficult to produce. Can open-source AI compete in the video generation space?.
