As many people know Chat GPT has taken the world by storm, allowing people for all professions to utilize it as a tool. From college students using it to proofread their papers to software engineers using it to code; and now not only generating images but videos from text.
While this has been in use for the last couple months the quality of these videos have been less than ideal. But now with the release of Sora on the near horizon, people can generate videos from text with relative accuracy.
Sora is different from other models produced in the past; one of the major improvements is that the video has a sense of object permanence. When one part of the frame is covered and then uncovered, the AI model remembers what was there instead of regenerating the uncovered section.
Previous versions of AI models that could generate videos were limited by the amount of computing power they had. Sora has much more computational power, which allows for smoother and more realistic videos. This also helps to cut down on the number of visual artifacts that occur.
With the soon-to-be-released Sora, the world will finally have the ability to generate videos from text with such accuracy that they could be nearly indistinguishable from the real thing. What that really means, we will all find out soon enough.