A company that makes AI that generates video from text. You write what you want to see and the program creates it frame by frame - already used in ads and film.
In practice it looks like this. You open a window, type something like 'a drone flies over the ocean at sunset, the camera slowly moves toward a cliff', and the program hands you a few seconds of moving video that looks shot, not animated. You can also start from a photo, bring it to life, change the background, add rain or fog to a frame you already have. This is not a filter on top of finished footage. Every frame is created from scratch, pixel by pixel, from your description.
It exists because a filmed shot costs money. Location, crew, lighting, days of editing - all of it is a budget that a small agency or an independent director rarely has. Runway takes the function of an entire editing suite, director and cinematographer combined, and shrinks it to a text field. It doesn't replace big cinema, but it replaces thousands of small productions for which a real shoot day was never an option.
Ad studios use it for quick visuals, when the client wants an idea in one day, not two weeks. Film crews use it too. Not to replace an actor - to save on an expensive special effect or test a scene before shooting it for real. The company was founded in New York in 2018 and, at its latest valuation, is worth over $5 billion, which shows how seriously the industry takes this direction.
Here's the catch. The result is still recognizable if you look closely. Hands, reflections, long movements - the physics sometimes breaks. For a five-second ad, no one will notice. For a two-hour feature film, that's a different question, and that's exactly why the company increasingly talks not just about video, but about models that understand how the world behaves.
Let me put it simply
I come from productions. I know how much a day on a film set costs, and how many ideas die because the budget doesn't stretch. A tool that gives you a rough but convincing shot from just a description isn't a toy - it's a change in who can even tell a story with pictures.
I don't like the hype around it, but the possibility itself is real. The next great visual storyteller might not have a camera, just a good sentence and the patience to fix it fifty times until it works.