AI is not just about writing texts or making pictures any more. Open AI has made a new tool (工具) called Sora. With just a sentence, it can make your words turn into a video (视频). The video looks good and closely follows what you want it to show.
Sora is better than many other tools that do similar things. It can make videos that fit different screens (屏幕), like big TVs or the screens on our phones. This means it can make many kinds of videos. It also lets people try out their ideas in a simple video before making abetter one, which makes things faster.
What's really cool about Sora is how it can make videos look real and smooth. It can show places with people, movements, and backgrounds, even changing the scene in the same video. Sora knows how things should look like in the real world. For example, if a woman walks by water, you can see her reflection (倒影) change.
Sora works so well because it has a helper named Chat GPT. Chat GPT can turn simple sentences into instructions, making the videos closer to what people want.
But Sora isn't perfect. It might feel puzzled with difficult movements or how things change after an action. For example, if someone bites (咬) a cookie in a video, the cookie might not look right. Sora might also mix up left and right or get the order of events wrong. This might happen because Sora learns about videos piece by piece and doesn't always know how they fit together.
Right now, Sora is still being tested. Open AI is asking people who know a lot about Internet safety to try and find problems with Sora to make it better. They've also let some artists and movie makers try Sora so they can give advice on how to improve it.