OpenAI unveiled its video generation model Sora two weeks ago. The technical report emphatically suggests that video generation models like Sora are world simulators. Are they? What does that even mean? I'm taking a deep dive into these questions in a new blog post.
https://artificialcognition.net/posts/video-generation-world-simulators/
(Grab a levitating cup of coffee, it's a bit of a long read)
Featuring statements & work by Tomer Ullman, Saining Xie, Peter Bataglia, David Ha, Jürgen Schmidhuber, Jim Fan, Yann LeCun, Jim Fan, Nando de Freitas, Neel Nanda, Kenneth Li & many others
This is my first real blog post after setting up the blog itself months ago, but I'll try to post somewhat regularly as a useful exercise to write up ideas at an early stage. Feedback/discussion on form or content is very much welcome!