← 返回首页
RT Shen Zhuoran: 1. Understanding and generating videos are mutually beneficial. They are the training materials for physics and embodied AI, e.g. Opt...
RT Shen Zhuoran
1. Understanding and generating videos are mutually beneficial. They are the training materials for physics and embodied AI, e.g. Optimus.
2. Real-time video generation is the most direct and flexible AI-native user interface. It's possible that it will subsume all frontend UIs.
𝕋𝕒𝕧𝕚: @elonmusk How so? What’s the goal with image generation, beyond creating content for social media?
1. Understanding and generating videos are mutually beneficial. They are the training materials for physics and embodied AI, e.g. Optimus.
2. Real-time video generation is the most direct and flexible AI-native user interface. It's possible that it will subsume all frontend UIs.
𝕋𝕒𝕧𝕚: @elonmusk How so? What’s the goal with image generation, beyond creating content for social media?