Nothing seems capable of stopping OpenAI entropy. The company specializing in the development of artificial intelligence had already managed to strike a big blow at the end of 2022. It then presented ChatGPT to the whole world. The language model has since become a real social phenomenon, used by millions of people every day.
But the objective forOpenAIwas never to limit ourselves to these textual responses. IfChatGPThas progressed since its release at the end of 2022, the language model remains limited in its creative process. Impossible for him to create images or videos.
It is to respond to this request,that OpenAI created Dall-E. This artificial intelligence can design unique images based on a text request. But today, the company led by Sam Altman wants to go even further. She just announcedthe arrival of Sora,“text to video” artificial intelligence.
In other words, Sora will be able to create a video of a few seconds from scratch, simply from a textual request from the user. According to Sam Altman himself, Sora would be capable ofproduce one-minute videos without loss of visual quality.
In order to demonstrate this technology, Sam Altman asked Internet users on X to write their ownprompt.He then integrated it into Sora and shared the video, much to the surprise of users. Indeed, creating videos from a simple request is a real headache. Other companies like Google have already tried to develop AI to respond to these queries but the result does not exceed a few seconds.
https://t.co/rPqToLo6J3 pic.twitter.com/nPPH2bP6IZ
— Sam Altman (@sama)February 15, 2024
Sora's big strong point is his ability to last over a distance. Artificial intelligence manages to produce a video but above all, to make it last. As OpenAI explains, Sora is capable of responding to the user's request but also of respecting all the fundamental laws of physics (gravity, force of attraction, fluid mechanics, etc.).
Sora is currently not available to the general public. Creating a video is far too energy intensive for anyone to use this system. Only a few researchers and selected content creators have access to Sora to demonstrate it.
You have to wait several hours for Sora to produce a video in photo-realistic quality lasting one minute. This is still far too much and OpenAI will work in the coming months to reduce this loading time as much as possible.
i-nfo.fr - Official iPhon.fr app
By : Keleops AG
