Just saw a 36Kr discussion, and the title was pretty blunt: “From Sora to Kling, video AI hasn’t reached its GPT moment yet.” As someone in the editing game, I’ve already felt it. When ChatGPT dropped for text, it literally changed how we work overnight—you type a few lines and it’s good to go. Video is different.
The tools are getting way more powerful, but we’re still far from “generate whatever you want, stable and controllable, plug-and-play.” Right now, it’s mostly used for B-roll, transitions, or concept clips. If you actually try to use a generated shot in a serious video, you’ll still run into issues like continuity, character consistency, and messed-up fingers or mouths.
I’m not into shouting “this is a game-changer” every time. The tools are still climbing; they help, but they’re not replacing anyone yet. The real moment will come when you can just say one sentence and get a finished clip that’s ready for approval.