Sora, openAI's latest Vincent video model, which was just released a few days ago, has been quickly noticed by "gold diggers". From the 17th, when Sora was widely discussed in China, to the 18th, in just one day, many paid tutorials related to Sora appeared on the Internet.
On the official website of openAI, we can see the sample video generated by Sora and a basic introduction to this model. Currently, Sora is not fully open to public beta. Only some visual artists, designers, filmmakers, openAI employees and other related people have obtained it. Sora's access rights, the constantly updated Sora-generated videos that the public currently sees are basically produced by these testers. However, openAI CEO Sam Altman takes orders online on the social platform. Netizens have written prompt words and basic scripts on the social platform @Sam Altman, and they may be replied, and Sam will give the corresponding script in Sora. generated video.
At present, the threshold for using Sora is still relatively high. On the one hand, it is very difficult to obtain a test account. On the other hand, in terms of usage, like large artificial intelligence models such as DELL and chatGPT, Sora also requires human input. Only prompt words and sentences can generate videos. How to correctly write sentences that can be understood by a large artificial intelligence model and train it to generate works close to the ideal state is not easy.
“Gold diggers” have been rapidly exploring this new land. On a certain paid knowledge course platform, the reporter found several paid courses related to Sora. One of the courses priced at 99 yuan stated in the introduction: The latest Sora first-hand information, first released on the entire network, be the first to eat fish!" There are paid tutorials and free tutorials, and a large number of free tutorials are rapidly appearing on the Internet to teach how to use Sora and how to register Sora, but this The content is very mixed, and the security of unknown links included in the tutorial is unknown. The situation is somewhat similar to what happened when chatGPT was launched at the beginning of last year. In addition to paying for knowledge courses, some people have also begun to focus on selling accounts and invitation codes, such as a small program named Sora, which was developed by an individual. The mini program changed its name to a name containing Sora on February 17, and launched a Sora list registration service. Others have caught up with the traffic of the Sora name. Although the name of the mini program is related to Sora, the content when you click on it is It was irrelevant, but after a closer look I found that the account name was changed on February 18th. It is conceivable that, judging from the path that chatGPT and DELL have gone through last year, there will be a lot of Sora package software in the future. After Sora opens to public beta, some emerging career directions may also emerge around Sora, such as professional Engaged in artificial intelligence script writing, a short video account specializing in playing videos generated by Sora. Sora combines e-commerce to promote videos. Of course, there may also be a group of cyber creators. The text created by AI is put into Sora to generate AI videos, and then used The AI-generated voice is used for dubbing, and the AI-generated music is paired with it to finally present a complete work.
Judging from the path that AI drawing has taken, it cannot be achieved overnight from a just-released conceptual product to one that can be more maturely used in business. On the other hand, there is still a considerable gap in the commercial application of AI-generated drawings at the social level. On the one hand, the controversy over the details is imperfect and distorted. Even if the industry has basically recognized the use of large AI models as tools, it is difficult for the public to accept such products. For example, the poster of the protagonist of the game "Beyond the World" not long ago The suspected use of AI has been met with fierce protests from players. The current shortcomings of Sora are obvious. This open model is still difficult to simulate the physical principles of complex scenes, and it cannot understand the causal relationship between objects. The spatial details are often confused, and it is difficult to accurately represent it. Judging from the current works presented Look, the effect is indeed explosive, but it is not enough to pose a huge threat to the traditional film and television industry, game industry and other industries closely related to video. Relevant industry giants also have different attitudes towards the Sora product. The game company Ubisoft believes this is a "huge leap"; Yin Ye, CEO of BGI Group, said "I would like to compare it to 'opening the Newtonian era of AI development'"; Meta's chief artificial intelligence scientist, deep learning pioneer, figure Spirit Award winner LeCun angrily criticized Sora for not understanding the physical world.
In any case, the relevant industries have already stood at a fork in the road of transformation. Huge controversies and subsequent social and ethical issues have also followed, and a new storm is brewing.
Yangzi Evening News/Ziniu News reporter Shen Zhao