Google just unveiled Gemini Omni, an AI model that can generate anything from any input — starting with video. You can type a description, upload an image, or even give audio, and it produces a realistic video output.
It’s physics‑aware (gravity, energy, fluid dynamics), so scenes look...