Google's Gemini Omni blurs the line between text prompt and video simulation
At I/O 2026, Google DeepMind unveiled Gemini Omni, a multimodal family that generates video from combined image, audio, and text inputs, signaling a shift from generative to simulational AI.