The true face of Sora! In depth good article .

#Sora exploded and sold classes to get rich first-

The discussion about SORA is probably the hottest topic in recent days. This is another bombshell in the technology field after OpenAI launched ChatGPT two years ago. As a generative model, you only need to enter a simple sentence, and Sora can help you generate an explosion effect for 60 seconds. From the point of view of scene effect, realism, and coherence, the ** generated by SORA is completely crushing to the rough ** generated by Runway, Pika, etc. on the market. For example, if you enter a woman walking in a bustling city at night, it is likely to come out like this.

An SUV is driving on a winding road, **yes.

There are also situations like sharks swimming among high-rise buildings, mammoths running on forests and snowfields, which are impossible to happen in reality, and Sora can do it. In the newly released generation ** works of the SORA team, there are even multi-angle camera positions. Some netizens said: "I didn't believe it was true when I heard about it, but I couldn't tell it was fake after **." ”

Zak Kukoff, an investor in San Francisco, publicly stated that in the next five years, an ordinary person can create a movie. Don't underestimate Sora's ability to "confuse the real with the fake". Last year, Oscar-winning actor Will Smith's ai** eating noodles became popular, and after Sora appeared, Will Smith posted another ** of his latest noodles eating, with the text: Generated by AI. But just when everyone exclaimed the ability of AI, Smith said that this was actually his real shooting of noodles**. To a certain extent, people are already a little confused about AI and reality. Many bigwigs are full of praise for the appearance of Sora. Elon Musk commented under one of Sora's dailies: "Humans are willing to gamble and lose". Zhou Hongyi, the founder of 360, has become a loyal fan of SORA, and he has issued six ** in a row, pointing out that SORA will subvert the film and television industry. ** has already reacted. On the second day of Sora's launch, Adobe, an image processing and production software tool company, had a stock price of more than 7%.

Shutterstock, another U.S. stock that sells ** and **, lost more than $70 million in market value overnight.

In Hollywood, which experienced the largest strike in history last year, the argument that human beings are being replaced by AI has spread again. But this does not affect the happiness of businessmen. As soon as SORA appeared, some netizens gave ideas for using SORA to make money, such as selling SORA's internal beta accounts, paying to make AI**, AI**Transfer**, selling courses to teach people how to use SORA, and developing SORA-related ** or tools, etc. Technician @Ai Doubi, he made a sora Wensheng by himself***, but what he didn't expect was that he was charged for cutting leeks when he turned his head. The hottest thing to make money with SORA is selling classes.

"SORA exclusive prompt word tutorial, limited time**999" and "SORA beta account, only 299".

Among them, the most famous is a Tsinghua doctor named Li Yizhou, known as "the first person in China's AI", along with Sam Altman, the founder of Open AI. Li Yizhou's AI class started last year, according to the number of store orders, his "Everyone's Artificial Intelligence Lesson", priced at 199 yuan, sold about 250,000 sets last year, about 50 million. He also built an AI platform called "Yizhou Intelligence", on which students can generate ** with text, and one requires 10,000 computing power. Buy the course and get "computing power", you need to charge the membership when you run out, **up to 399 yuan per month. No wonder some people say: In China, the first wave of successful use of AI to make money is the tutor who sells AI courses. However, here's a basic fact: OpenAI is currently only providing SORA access primarily to a select group of people, such as designers and filmmakers, and is not being released to the public. Therefore, soon someone scolded Li Yizhou for cutting leeks and began to fake his course:

The quality of the course is very watery, "Yizhou Intelligence" is suspected of being a shell chatgpt, not Tsinghua University's science and engineering but art...

But one Li Yizhou fell, and countless people who used sara to cut leeks got up. According to the consumer service platform "Black Cat Complaints", there are currently nearly 3,000 complaints related to AI courses, and most of the complaints are due to false propaganda and low quality of courses. In the end, even the official people of openai came out to refute the rumors: there is no permission to open to the outside world, don't be deceived.

#The creativity of the Chinese Internet is all used to make money-

There is a miraculous phenomenon that in the more than 20 years of China's Internet, every time a new technology or new concept comes out, almost all of the first people to make money are selling courses. In the past two years, ChatGPT, the metaverse, and the blockchain in the past few years have all seen such things. If you happen to have experienced the rapid development stage of China's mobile Internet in the past 10 years, most of you have a deep memory of a phenomenon around 2014. In those years, in the café in Zhongguancun, Beijing, college graduates could be seen everywhere holding PPTs made overnight, talking about their entrepreneurial projects with emotion, and in the whole process, a large number of novel words such as "O2O", "business model", and "wool from pigs" blurted out from their mouths, making people dizzy to hear. But such a speech has a great possibility of getting millions of angel investments. This is a true microcosm of the tide of Internet entrepreneurship in China back then, with everyone opening their mouths and closing their mouths about business models, and concepts such as "Internet thinking", "cross-border" and "closed-loop" flying all over the sky. Many people find that a migrant worker in the workplace who was just like themselves two years ago has suddenly become the boss of an Internet start-up. It was that crazy period that allowed an unwritten "iron law" to spread in the Internet industry - there are not enough products, and the concept is to make up. Whoever can tell a good story and play with the "concept" will be able to become a pig on that tuyere and grasp the wealth code of that era. In the following ten years, this "wealth password" has been iterated and updated several times, becoming "sharing", "autonomous driving", "blockchain", "metaverse", and until now ChatGPT, SORA, and AIGC. Two years ago, the concept of "metaverse" became popular, and you may still remember it. A course called "The First Lesson of the Metaverse", although the course content is only some basic industry reports and industry books, dares to price 688 yuan. It has only been online for 10 days, and the cumulative revenue is nearly 1.6 million.

Many people still remember the slogan on the poster of that course: "688 at a time, ten years of chasing the metaverse". Immediately afterwards, the concept of "metaverse real estate" flew again, and people who were bored in real housing prices began to plan their own "digital real estate", and the three "virtual land" that Lin Junjie spent millions to buy once became the most popular topic at that time.

What is happening today in the field of AI is very similar to the metaverse. Register an account with "AI" on the short** platform, write a service introduction similar to "AI Consulting", and even if you don't post works, you can receive a lot of private messages. When ChatGPT first came out, a blogger on the Knowledge Planet quickly established a community, priced at 288 yuan, and tens of thousands of people signed up within a few days. However, some people soon regretted it, because in addition to being able to get a ChatGPT experience account with a giant card, the rest of what they could gain was the chatgpt industry dynamics updated daily by bloggers. You know, if you can search on the Internet, you can find these dynamics. The layout of various enterprises and large manufacturers in the AI field during this period can also illustrate the problem from another side. Launched the generative dialogue product "Wenxin Yiyan", which has more than 100 million users in just one year; Meituan co-founder Wang Huiwen founded the AI startup "Light Years Away", which was selected by Meituan for 20$6.5 billion ** acquisition; Wang Xiaochuan, the former CEO of Sogou, has completed $300 million in financing in just half a year since its establishment, with Alibaba, Tencent, Xiaomi, etc. all participating. In addition, Xiaomi released the first AI large model computational photography platform Xiaomi AISP, and OPPO has officially entered the era of AI mobile phones, etc. When most people are still in a state of ignorance of "what is AI", the Internet bigwigs have regarded you as one of the future "users" and started a new round of capital expansion. In 1995, when the Internet was just emerging around the world, the prophet Kevin Kelly said, "The good news is that you're going to be a millionaire; The bad news is that everyone becomes a millionaire. ”

But it has been proven time and again that Kevin Kelly may be overly optimistic. I thought that the Internet, which was born with genes such as "inclusive and inclusive" and "beneficial to the people", ended up allowing a very small number of businessmen with a keen sense of smell to quickly accumulate huge wealth in a short period of time. It's like Sora, I don't know if it's Pandora's box, but it must be the magic lamp of the class seller. #AI anxiety with nowhere to put it.

In fact, in the final analysis, the popularity of selling courses reflects the public's urgent need for AI to understand, and behind becoming a "leek", it is nothing more than people's anxiety about AI, worried that they will not understand it and be replaced. Film and television practitioners are probably the most anxious group of people. Not only the post-production software production market like Adobe, but also front-end science fiction movies, 3D games and other related fields will most likely be the first to usher in the commercialization impact of SORA. Wang Ran, CEO of Yikai Capital, said that he thought that the first to be replaced by AI were those repetitive mechanical jobs, such as blue-collar workers and workers, and creative workers were the most difficult to replace, but the truth was the opposite. In an article on Phoenix.com "Recording 42 Chinese Impacted by Sora", a storyboard named Qiuller just pointed out under a ** that "AI can't solve the problem of picture coherence" and breathed a sigh of relief, 2 days later, Sora appeared. A group of netizens who watched the excitement ran to the bottom of his comment and ridiculed: "Sora is already here, tremble." ”

According to the "2023 Employment Trend Data Report" from Liepin Big Data Research Institute, compared with 2022, the number of jobs requiring mastery of AIGC - "AI-generated content" increased by nearly 180% year-on-year. This kind of anxiety in the industry is best known to the sellers. People need professional explanations, so Li Yizhou labeled them as "Tsinghua Doctors", and people wanted to use AI to make money, so Li Yizhou wrote in the introduction of the product that a student had used AI to earn more than 10,000 yuan a month. Moms and newcomers in the workplace, who account for a large proportion of the paid staff, signed up for the course with great expectations, but found that the higher the expectations, the greater the disappointment, and the professionals who want to use AI to improve their work skills seem to understand, but find that there is nowhere to get started. Want in-depth service at your fingertips? Either there is none, or private training 3999. In fact, as long as you calm down, you will understand a basic truth - "anxiety", which means that things have not happened yet.

Regarding the current SORA, the OpenAI team publicly said: "At present, SORA generation may contain illogical images, confusing spatial details such as left and right, and it is difficult to accurately simulate the physical principles and causal cases of complex scenes." For example, in the ** generated by SORA about "people want cookies", it is likely that there are no bite marks. This can be explained by the fact that it is easy to feed the AI the world's images, but it is much harder for the AI to understand why things in the picture exist in the way they do.

It's like the alpha dog that once beat a human Go player, but the logic of its chess play is completely different from that of humans. Lee Sedol, who lost to Alpha Dog that year, said: Some of Alpha Dog's moves seem to be illogical. Back to Sora, if it is open for use one day, there will definitely be a large number of "cookies without bite marks" and various logical loopholes**. What's more, the application of a new technology is often trapped by factors other than technology. In order to test the safety performance of the vehicle, Google's autonomous driving team recruited a group of volunteers internally, and told the volunteers that the vehicle performance was not perfect, and it was still necessary to maintain a driving posture and respond to emergencies at any time. But it turned out that no one listened. Because under normal road conditions, the vehicle drives automatically very well, which makes the volunteers feel very safe, so they simply leave it all to the automatic driving, and they should watch**see**, and they should sleep. This made the Google team understand that once the self-driving technology reached a high enough level, passengers would take it for granted that they would give up driving. At this time, regardless of whether the software is risky or not, whether the road conditions will encounter unexpected situations, the driver will not maintain 100% attention.

Source@motor trendsora will also encounter factors outside of technology, as some people have begun to worry: "The real world does not exist anymore." "In this case, it may be too early to say large-scale commercialization or disruption of the industry. The actions of several foreign capital tycoons also seem to be a signal. For example, Musk, who has always been ahead of all mankind, has stepped down from the board of directors of Open AI. Meta Chief Scientist and Turing Award winner Yann Lecun also publicly expressed his disapproval of SORA at the world** summit a few days ago

"Let me clear up a huge misconception that generating most of the things that look real from a prompt doesn't mean that the system really understands the physical world. Generating a future from a world model is not the same thing. The possibility of trustworthiness is huge, and the generation system only needs to output one sample to be successful. However, there is much less room for a reasonable possibility of continuation in reality, especially given a given action scenario. ”