Amy Wu’s Post

In the video AI space, 5 second heavily post-production edited shorts are impressive but aren’t close to truly disrupting how we watch and create videos today. The devil will be in control and realism, maintaining character-consistency, maintaining storyline-consistency, crossing the uncanny valley of physics realism, and being able to create much longer video content. Text-to-video is not going to be the solution, but rather agentic image-to-video creation and newer techniques. As we move to longer form video, as top AI video companies open their APIs (Kling, Minimax, Sora etc) the old truism in media will prevail: “content is king.” It will be as important to have talented creatives and writers storyboard and know how to create the best stories from the near-term limitations of gen video as building the video production itself. The space is moving at lightning fast speed. Excited to see our portco Higgsfield AI launch ReelMagic today in a major step producing long-form content and to follow where they take this. 🎬🎥 https://lnkd.in/gjPjCHjp