Sora is here (openai.com)

Article summary

The article appears to be about the release of Sora, a new AI video generation tool from OpenAI, although the article itself is not available. The discussion reveals that Sora is part of the ChatGPT subscription and has limitations such as watermarking and video length restrictions at the lower pricing tier. The tool's capabilities and pricing are compared to other open-source video models. The release of Sora has sparked a discussion about the current state of AI video generation and its potential applications.

Main themes

• AI video generation
• OpenAI Sora
• Pricing and limitations
• Open-source alternatives
• Technical challenges
• Industry applications

What commenters say

• The release of Sora is seen as a marketing promotion rather than a significant improvement in AI video generation capabilities.
• Open-source video models, such as Hunyuan, are considered to be more capable and cost-effective than Sora.
• The current state of AI video generation is limited by technical issues such as object permanence, temporal stability, and physics comprehension.
• The pricing and limitations of Sora, including watermarking and video length restrictions, are seen as restrictive and potentially uncompetitive.
• Some commenters believe that AI video generation is not yet ready for industry applications such as Hollywood, while others see potential uses in advertising and social media.
• The need for human-guided edits and the low percentage of usable output are major challenges for AI video generation startups.
• Incorporating physical constraints, such as Hamiltonian constraints, could potentially improve the quality of AI-generated videos, but this is seen as a difficult task.
• The quality of AI-generated videos is hindered by motion artifacts, particularly in scenes involving fluid or dynamic movements.