The news of October 10thOpenAI Sora Bill Peebles, who was in charge yesterday, 9 October, issued a tweet on platform X announcing that Sora had been on line for less than five days and that the number of downloads had exceeded 1 million times, a milestone that was even faster than the phenomenon-class product under the corporate flag ChatGPT.
1AI QUOTES BOVEN THAT THIS DATA WAS ACHIEVED WHEN THE APPLICATION WAS STILL IN THE INVITATION PHASE AND WAS OPEN ONLY TO USERS IN THE UNITED STATES AND CANADASora is now number one on the Apple Store free application list。
Sora application, published at the end of September, can be considered an "AI version of the shivering" with the core function of enabling users to generate new video content using their own or friends' real video clips as material。
Users need to upload a short film called “Cameo”, which can then be used as the basis for AI creation. At the same time, there is a sharing function in the application, which allows users to choose to share their "cameo" with a specific group or all Sora users, allowing others to use their own image for AI video generation。
At the technical level, Sora relies on the powerful Sora 2 video generation model of OpenAI. The model is capable of processing complex dynamic images, supporting the generation of a variety of styles, including writing, film senses and animation, and has the capability to synchronize the generation of audio, ensuring that all created videos are audio-based and synchronized。
In product design, OpenAI emphasized Sora ' s social attributes, stating that they were intended for interaction and creativity among friends, which is why it adopted an invitation system. The information stream in the application displays content on the basis of the object of interest or interaction of the user and regularly asks how the user feels about its use。