（2023-04-06）
Since ChatGPT became popular, it has set off a new boom in the big A market. The concept of AI has become a new outlet beyond new energy, rising in turn, and the so-called new energy dogs do not buy. There are even institutions that secretly change positions to buy AI and stage sheep's head dog meat.
However, one thing has affected the continued rise of the concept of AI. OpenAI, the parent company of ChatGPT, announced that it began to stop subscribing to the service due to computing power limitations, and the latest news is that subscriptions have been restored.
Before this, there was a more suffocating news, OpenAI officials began to block a large area from March 30, prohibiting the use of Asian nodes to log in, and some netizens also revealed that OpenAI has suspended the registration of new accounts.
Obviously, ChatGPT is widely used by Asian users to take some measures to protect their own interests. But judging by the effect, it is clearly unsatisfactory.
In fact, my first contact and trial of ChatGPT was in a WeChat group, and the model at that time was that domestic enterprises transferred ChatGPT to the WeChat group through their own technology to carry out channel services. At the same time, it is also peddling access service members.
Just as you build roads, I build bridges, each has its own way. After OpenAI built ChatGPT, an AI avenue, some domestic companies began to build bridges on the AI avenue, borrowing ChatGPT to start charging bridge fees.
The old rent-collecting business has gained new vitality in the AI era.
However, in the process of taking off the concept of AI in this round of big A market, there is such a news that attracts a topic.
On April 3, there was news on the Internet that ChatGPT banned a large number of Asian accounts, and Tomcat's product development may be affected. In response, Tomcat responded that the relevant news was untrue, and the company's API access is normal.
Since it is API access, then Tom Cat is a member of the majority of bridge fees, and then look at this Tom Cat, he is still a listed company, and claims to be self-developing his own AI.
It follows from this that this will not be an isolated case. There are many companies that collect bridge fees by accessing ChatGPT, and after the AI frenzy, the market will gradually return to the normal track.
After all, after the development of things, the market will also see clearly that it is computing power and electricity that restrict the development of AI. The core of computing power is chip semiconductors, so semiconductors replace AI to become a new hot spot in the big A market.
After the market returns, the future of AI also returns. The domestic AI company that collects bridge fees has made OpenAI's official borrowing of computing power issue start to ban and restrict subscriptions through large-scale computing power behavior.
It can be said that the domestic wool enterprises staged a burlesque of punching the teacher to death. After all, the cost of computing power construction investment is there, and if the other side is weaker, our side is naturally stronger.
So in the future development of AI, which one will laugh at the rivers and lakes?
According to Science Network reports, in the 2022 World Artificial Intelligence Computing Power Ranking, Pengcheng Cloud Brain II continues to top the list.
The list was released at the China Super Computing Power Conference, jointly released by the Institute of Computer Science of the Chinese Academy of Sciences and the Argonne Public Laboratory in the United States, and the Zurich Laboratory in Switzerland. In terms of professionalism and authority, it is the most worthy of reference.
Obviously, in terms of computing power, China is in the leading position in the world. And the other big head, electricity, is also a leading position in the country.
According to data from China's power grid, China was still the country with the largest electricity production and the largest increase in the world in 2019. China's power generation reached 7503.4TWh in 2019, an increase of 4.7% over 2018, and China's total power generation accounted for 27.8% of the world total; The United States, Japan, Germany, France and the United Kingdom all declined.
Power generation in major countries in 2014~2019 (unit: TWh)
The 20 countries with the highest electricity consumption in 2014~2019
From AI computing power to power supply, it is clear that the future of AI is still in China. The current gap is only in the training of large models. As I said in a previous article, the gap between domestic AI manufacturers and international manufacturers is not large.
OpenAI's leading position is not its computing power and algorithms, but the first-mover advantage of ChatGPT. With this lead and more than 100 million users, ChatGPT is in the first echelon of AI training, and this trend is not unsurpassable.
On the other hand, Europe has begun the traditional suppression of American technology Internet companies. On March 31, the Italian Personal Data Protection Agency announced that it will ban the use of chatbot ChatGPT from now on, restrict its development company OpenAI from processing Italian user information, and open an investigation. Subsequently, Germany, France, Ireland, Spain, Canada, etc. will also join.
Obviously, behind the EU's close pace with the United States, countries still have their own small nine-nines.
As long as the domestic big factories can stabilize their performance. With the accumulation of domestic computing power and electricity, domestic AI is still promising.
Isn't it thanks to the company that makes ChatGPT wool?