OpenAI launches Sora: instant videos out of text prompts

Why is Sora a leap forward in the AI world?

In the past we have seen plenty of text to video tools, tech giants such as Google and Meta have shared similar technology in the past but none of them can capture the sheer quality of Sora. In terms of producing quality content, Sora seems to be not only a few steps but miles away from what was considered normal or capable.

OpenAI has taken the quality aspect of artificial generative content to a height that was deemed impossible but they have done it with ChatGPT now, it’s time for the videos as well.

Everything you see in the video crafted just from words, the quality to every little thing from the snow, tree, clothes, hair, building, sky, and more is dipped in high-definition quality. The potential Sora has to offer is mind-numbing. The video has garnered 35 million views in just 11 hours. This goes on to tell you how Sora is breaking new grounds in the AI world.

Here is why Sora is a massive to the AI community, this here features the list of things Sora can do, almost instantly –

Sora model carries in-depth understanding of the language, as it interprets the data accurately to create the video instantly. The range of interpretation is what sets Sora apart from others, as it can also take text prompts or input’s character expression, motion into account. Nearly every aspect of the text that gets written down, Sora takes every little word of the text seriously to craft a video that doesn’t miss out on anything.

Limitation and safety of Sora

There is one thing we know about AI, it is a long road to perfection, Sora is no different. The tool is nowhere near the perfect stage, it still has a long way to go to reach that point. Some errors are still present in Sora; a statement released by the OpenAI has said Sora might confuse the spatial details of the prompt and will have some difficulty in following the camera trajectory.

The AI might struggle with creating physics in more complex scenes and might have a difficult time understanding the cause and effect. A longer text input for a video might not lead to good results. Sora can definitely create videos out of text, but if the text goes on to create more and more complex scenarios, then Sora might fail to deliver that to the surface.

The company is also pushing safety to the AI by working on terms such as misinformation, hateful content, and bias. The company is also planning to release a detector tool that can identify the AI video from the real one. OpenAI also said to make the output video adhere to the user policies, as they have the power to review every frame of the video to ensure it stays in the circle of their policies.