Sand AI has introduced its Magi-1 model, an openly licensed video-generating AI that has received acclaim from leaders in technology, such as Kai-Fu Lee of Microsoft Research Asia. However, testing by TechCrunch found that Sand AI actively censors certain images from being uploaded, such as those depicting Xi Jinping and Tiananmen Square, suggesting compliance with Chinese regulatory standards. While Magi-1 boasts advanced video creation capabilities, its hefty hardware requirements mean that most users rely on Sand AI's platform to experiment with this innovative technology.
Magi-1, developed by Sand AI, effectively generates high-quality videos, but the model includes stringent censorship against politically sensitive content, limiting user input.
The new model, Magi-1, is praised for its ability to generate videos with impressive realism, although it requires extensive hardware resources for operation.
Collection
[
|
...
]