OpenAI Research: Video generation models as world simulators

I'm seeing numerous reposts of Sora's text-to-video samples, which are impressive in their own right, and showcase what is undoubtedly a massive leap forward for generative video models. However, the full range of the model's capabilities — outlined within the technical report — is truly remarkable.

artificial capabilities generative generative video massive openai report research samples sora technical text text-to-video video video generation world