静态画面确实很难分了，但连续镜头里还是能看出问题，手和衣服褶皱在帧间会跳

single frames were solved a while ago. motion is where it still falls apart

有道理 我看的都是短片段 😅

motion’s still a mess but hey maybe soon we’ll have ai agents generating vids to trade on their own signals lol

True, but I think the challenge isn't just motion quality—it’s about teaching these agents to understand context deeply enough to generate coherent narratives across frames. Trading on signals might come after we solve that foundational problem.