AI Applied: Covering AI News, Interviews and Tools - ChatGPT, Midjourney, Gemini, OpenAI, Anthropic Sora's Launch: A New Era for AI Video
Dec 20, 2024
Excitement abounds as OpenAI's Sora launches, but it comes with critiques about performance and accessibility. The discussion outlines the cost structure for creators, especially for high-definition content. They dive into the strengths and weaknesses of Sora’s video technology, comparing it to competitors. Usability for both casual users and large companies is examined, alongside a nod to innovative applications in advertising. The future of generative AI in video production looks bright, with 2025 set as a key milestone.
AI Snips
Chapters
Transcript
Episode notes
Sora's Physics Critique
- Marcus Brownlee critiqued Sora's physics, showing a Frisbee morphing into odd shapes.
- OpenAI's Sora, while impressive, doesn't perfectly handle object physics in videos.
Sora's Strengths
- Sora excels at parsing different video elements, like a dog, balloon, and museum, distinctly.
- It also renders realistic features like shadows, improving overall video quality.
Sora's Cost Structure
- Sora uses a credit system; ChatGPT+ gets 1,000, Pro 10,000.
- 1080p video costs 200 credits for 5 seconds, limiting output even for Pro users.