Automating Invisible Video Editing Labor

Diving deeper into

Cristóbal Valenzuela, CEO of Runway, on rethinking the primitives of video

Interview
Some estimates say that around 80% of the time spent on video editing involves uncreative operations such as tedious frame-by-frame editing.
Analyzed 5 sources

The real opportunity in AI video is not replacing editors, it is deleting the invisible labor that makes video slow and expensive. In practice, that means automating jobs like rotoscoping, object removal, transcription, subtitle generation, and frame level cleanup, the work that turns one creative decision into hundreds of repetitive clicks. Runway built around that bottleneck first, which is why it could move from a model playground into a production tool for marketers, creators, and VFX teams.

  • The clearest early wedge was rotoscoping. In traditional workflows, an editor or VFX artist traces or masks subjects across many frames by hand. Runway turned that into a model driven task, and later added inpainting, depth estimation, noise removal, and transcription, all aimed at compressing hours of cleanup into minutes.
  • This is where Runway differs from tools like Descript and Veed. Descript simplifies editing by letting users work through transcripts, and Veed bundles browser based editing with AI helpers like captions and background removal. Runway goes deeper into pixel level and shot level automation, where the user is changing the video itself, not just trimming, captioning, or repackaging it.
  • That workflow focus also explains the business model. Instead of selling a high end VFX seat like Nuke, or just bundling lightweight video into a broader suite like Canva, Runway packages expensive automation into a subscription product that can cut per shot costs sharply for small teams. That makes professional post production usable by far more people inside a company, not just trained editors.

The next step is that editing and generation merge into one workflow. As Runway keeps automating cleanup work and adds more native scene generation, camera control, and collaboration, video software moves away from timeline craftsmanship and toward directing outcomes, which expands the market from editors to anyone who needs to ship polished video quickly.