Generative Video at the Speed of Light — Keegan McCallum, uRun
Keegan McCallum from uRun presents how generative video has become much cheaper and faster — ten dollars now provides approximately three hours of continuous video.
He argues that quality is no longer the interesting factor, but rather the ability to control the video in real time. This opens entirely new use cases: a webcam showing the hairstyle you're considering, or tools for people who don't think in text. The major technical challenge is no longer creating the video but serving it globally and keeping it synchronized with the user's controls frame by frame.
Open on YouTube →
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
