First of all, the SORA model is not the first to make AI, there have been Stable Diffusion Video and Pika, as well as other companies that have launched their own products. SORA's explosion is mainly due to its own strong backend and OpenAI's convincing endorsement.
Sora detonated the second circle in communication, the first circle is the people who eat crabs, and then the people who eat the crabs are waiting for the next crab - Sora is the second circle, based on the popularity of the topic pioneers, further breaking the circle, advancing from AI enthusiasts to more industries.
If we are really curious about what will be the next hot spot, then we can find the most powerful and deepest companies or teams in some vertical fields, and then the works of these people will point out the direction of the next hot spot.
But as an ordinary person, or an ordinary student, a working person based on content production, or an ordinary person who just comes to have fun. This sora only served as a reminder: it turned out that there was such a thing in the world that I had not noticed in the first place, and its influence and its subversive power broke through the information cocoon and came to us. As information, he is very penetrating, so it is worth trying to understand this thing.
SORA, behind it is OpenAI, a role like the "Apple" of the artificial intelligence world. Initially, a language conversation model was introduced, and through some knowledge base, the model could understand human knowledge and questions, so as to answer them - this dialogue model is very novel, it is endless question and answer, until one party gets bored or the other party has a system failure (I don't think this can be understood in traditional question answering, and then we can discuss it in more detail), and then similar models are also studied in major companies, and various products are born (more like imitation).
Artificial intelligence understands the text, the next step is to understand** and sound, based on the diffusion model (no need to understand for the time being), you can learn ** or sound very well, and then two models of midjourney and stable diffusion were born respectively (openai also has a dalle series in this field), not that other products are not important, but the core products are these three, which can be made according to the language**.
It is worth mentioning that diffusion can do sound alone, but it does not have its own application scenarios, and can only play the role of dubbing in the ** production software. Then, after completing the breakthrough of text and voice, AI giants began to think about how to overcome the difficulties of production. **It can be understood as a combination of frame by frame** + sound, so the basic mode is the diffusion model, which is obtained by making frame by frame**.
Take the lead in pika and win the investment. Subsequently, Stable Diffusion also launched **production, Stable Diffusion Video (SDV), at this time, the production effect was further improved and a few months later, to this day, SORA has also launched, and the comprehensive quality of its demo** is better than that of the other two**, so it has been widely acclaimed. However, based on the common diffusion model kernel, it is still essentially an AI that relies on training data and prompt word skills, and the effect is high and low.
If you are willing to enter the circle under this wave of SORA, welcome to pay attention to this account and continue to output the dynamics and thoughts in the AI field.