VOID

Netflix’s AI team has open-sourced VOID, a model that solves one of video editing’s nastiest problems. Remove a person holding a guitar, and most tools leave you with a floating instrument that defies gravity. VOID doesn’t just erase objects, it reconstructs the physics of what should be there.

The model understands how light, shadows, and motion should behave after an object disappears. It fills gaps with contextually appropriate backgrounds and adjusts lighting to match the scene. This is the kind of work that takes Hollywood VFX teams weeks to get right.

VOID represents a shift from basic object removal to physics-aware scene reconstruction. It’s trained to understand how the world works, not just how to paint over things. For anyone working with video content, this could eliminate hours of tedious cleanup work.

The open-source release puts professional-grade video editing capabilities into more hands. We expect to see this integrated into editing workflows quickly, especially for content creators who need clean results without massive budgets.