It has the ability to generate high-definition videos lasting up to 1 minute based on textual commands input by users. These videos can encompass complex scenarios involving multiple characters, specific types of movements, precise thematic and background details, and more.
Sora not only accurately portrays details but also comprehends the existence of objects in the real world, generating corresponding characters and information.
In other words, Sora can understand the fundamental principles of our physical world. For example, when a snowflake falls onto a leaf, due to the force of gravity on Earth, the snowflake naturally slides down from the leaf. Similarly, if too much snow accumulates on a tree, it will bend the branches…
These common knowledge and rules in our natural world were previously beyond the understanding of AI. However, this time Sora unexpectedly grasps these natural laws, which is the remarkable aspect of Sora.
With this basic understanding of physics, the videos generated by AI become even more aligned with the realities of the natural world.
The result is that when watching videos generated by Sora, they closely resemble those captured in the natural world. To the extent that it may be difficult for the naked eye to distinguish between those generated by AI and those from nature.
It is this capability that has led many to recognize its potential for numerous commercial applications. Therefore, many believe that Sora could bring significant disruptions to industries such as advertising, movie trailers, and short video production. Perhaps soon, we will see scenes and segments in movies generated by Sora AI. What’s frightening is that when we watch them, we may not necessarily perceive or realize which segments are generated by AI and which are captured naturally by humans.
The boundary between reality and virtuality is becoming increasingly blurred.