Preview of OpenAI’s new model Sora
- OpenAI revealed a new model called Sora that can create hyper-realistic, minute-long videos from simple text prompts.
- The videos show a complex understanding of physics, camera movement, and storytelling.
- It’s not yet publicly available, but the examples have gone viral, raising both excitement and concern.
The AI world was set ablaze when OpenAI dropped the first demo videos from Sora. From a stylish woman walking down a rainy Tokyo street to a realistic shot of golden retriever puppies, the quality was leaps and bounds beyond previous video generators.
The model’s ability to maintain visual consistency and follow a prompt’s “story” is what’s most impressive. However, it also immediately sparked serious conversations about the ethical implications of creating such convincing deepfakes and the future of creative industries.
Source: OpenAI Blog – “Sora: Creating video from text”
