Keep pulling the thread on Bill Peebles, Rohan Sahai & Thomas Dimson.
Approximately 70% of Sora users are also creators on the platform.
OpenAI has announced it is beginning to introduce monetization features to the Sora app.
OpenAI has launched an API for its Sora video generation model.
OpenAI believes the next major breakthrough for video models will be the ability to simulate long-duration processes, enabling applications in scientific research for fields like biology and physics.
An OpenAI researcher predicts that state-of-the-art video models will become an essential tool for conducting research in biology, robotics, and material sciences.
An OpenAI researcher predicts the first major scientific breakthrough from video model simulation will likely be in the field of classical physics, such as modeling turbulence.
An OpenAI researcher would be "shocked" if video models are not enabling scientific breakthroughs by early 2028.
The inference cost for Sora 2 is orders of magnitude lower than the ~$50 per video cost of Sora 1, representing a significant decrease in cost alongside an increase in model intelligence.
The Sora app was number one in the App Store for approximately one month following its launch.
The entire AI industry, including OpenAI, is currently experiencing a shortage of compute resources.
Launching the compute-intensive Sora video app required reallocating GPU resources from other departments within OpenAI.
OpenAI internally prototyped and tested a version of ChatGPT that included a social media feed before launching its image generation features.