December 4, local time.OpenAI CEO AltmanIt was revealed on social media that each following weekday there will be alive streamingIn total, there were 12 live streams, which included launches or demos, including some big events and some Christmas presents.
According to The Verge, sources familiar with OpenAI's plans say thatThese new products include the highly anticipated text-to-video tool Sora and a new inference model.
The first live stream will take place tomorrow, but the exact release has not been officially confirmed. However, the Wall Street Journal has also previously reported that Sora could be released before the end of 2024.
1AI understands that Sora was first unveiled earlier this year and shared with a small group of beta testers. But according to the Washington Post, about 20 artists leaked the model to the public in protest of "unpaid labor." OpenAI has been widely criticized for the data sources it relies on for its language models, and Sora is no exception. The company has not directly addressed whether Sora was trained using public YouTube videos, which the YouTube CEO has said would violate the platform's terms of service.
OpenAI isn't the only company working on text-to-video modeling; Google has also unveiled its latest generative AI video model, Veo, which originally launched in May, three months after OpenAI announced Sora, and is now available for private preview through Google's Vertex AI platform.