As Chinese ‘ai heroes’ prevailed us to refrain from stunning the silicon valley

When Chatgpt broke into the world of artificial intelligence (AI), the inevitable question followed: did this write problems for China, the biggest American rival?

Two years later, the new AI model from China shifted the question: can I now stop the Chinese innovation?

For a while, Beijing seemed to have dropped the response to Chatgpt, which is not available in China.

Unpressed users mocked Ernie, Chatbot by Giant Baid for search engines. This was followed by versions of Technological companies Tencent and Byttenten, which were rejected as followers of Chatgpt – but not so good.

Washington was convinced that it was ahead and wanted to keep it that way. Thus, bidet administration increased limitations that prohibit exports of advanced chips and technologies in China.

That is why Deepsek’s launch has amazed the Silicon Valley and the world. The company says its powerful model is far cheaper than billion US companies that spent on AI.

So how is the little famous society – whose founder is said to the Chinese social media as “Ai hero” – pulled out this?

When the US banned the world’s leading manufacturers like Nvidia to sell advanced technology at China, it was certainly a blow.

These chips are key to the construction of powerful AI models that can perform a series of human tasks, from answers to basic inquiries to solving complex mathematical problems.

The founder of Deepseek Liang Weenfeng described the ban on the chip as his “main challenge” in interviews with local media.

Long before the ban, Deepseek acquired the “significant stock” of the NVIDIA A100 chips – estimates ranging from 10,000 to 50,000 – according to MIT Technology Review.

The leading AI models in the West use about 16,000 specialized chips. But Deepseek says he trained his AI model using 2,000 such chips and thousands of lower -class chips – which is what makes his product cheaper.

Some, including American technology billionaire Elon Musk, have questioned this claim, claiming that the company cannot reveal how many advanced chips are really benefits with respect to restrictions.

But experts say the ban of Washington has also brought the challenges and capabilities of the Chinese AI industry.

This was “forced by Chinese companies like Deepseek to innovate” so I can do more with less, says Marina Zhang, an associate professor at Sydney University.

“Although these restrictions are challenges, they have also encouraged creativity and resistance, aligning with the broadly Chinese political goals of achieving technological independence.”

The second largest economy in the world has largely invested in great technology – from batteries powered by electric vehicles and solar panels, to AI.

China’s conversion to technological superpower has long been the ambition of President XI Jinping, so Washington’s limitations were also a challenge that Beijing took over.

The publication of the new Deepseek model on January 20, when Donald Trump swore as an American president, was intentionally, according to Gregory C Allen, a professional for AI from the Center for Strategic and International Studies.

“The time and way of being said – this is exactly what the Chinese government wants everyone to think – that export controls do not function and that America is not a global leader in AI,” says Mr. Allen, former director of the Strategy and Policy in the US Ministry of Defense Common artificial intelligence center.

In recent years, the Chinese government has nurtured AI talent, offering scholarships and research scholarships and encouraging partnerships between the University and the industry.

The National Engineering Laboratory for Deep Learning and other initiatives supported by countries have helped to train thousands of AI experts, according to Ms. Zhang.

And China had a lot of bright employment engineers.

For example, take Deepseek’s team – Chinese media say that less than 140 people, most of whom have proudly proclaimed the “domestic talents” from elite Chinese universities.

Western observers missed the appearance of “new generations of entrepreneurs who give priority to based research and long -term technological progress during fast profit,” says Ms. Zhang.

The best Chinese universities create a “fast growing talent AI pool”, where even managers are often under 35 years of age.

“Growing up during the Chinese fast technological ascent, they are deeply motivated by the urge to confidence in innovation,” she adds.

The founder of the Deepseek Liang Wenfeng is an example of this-40-year-old studied AI at the prestigious Zhejiang University. In the article on the Tech Outlet 36K, people are introduced to him that “he is more like Geek, not the boss.”

And the Chinese media describes him as a “technical idealist” – he insists that Deepseek be an open code platform. In fact, experts also believe that the successful open code culture enabled the young starts to have their resources and progress faster.

Unlike larger Chinese technological companies, Deepsek has prioritized priority to explore, which has made more experimentation, according to experts and people who worked at the company.

“The first 50 talents in this field may not be in China, but we can build such people here,” Mr. Liang said in an interview with 36kr.

But experts wonder how further Deepseek can go. Mrs. Zhang says “new US limitations can limit access to US users data, which potentially affects how Chinese models like Depseek can become global.”

And others say that now they still have a huge advantage, such as Mr. Allen’s words, “their huge amounts of computer resources” – and it is also unclear how Deepseek will continue to use advanced chips to continue to improve the model.

But for now, Deepseek is enjoying their moment in the sun, given that most people in China have never heard of it until this weekend.

His sudden glory saw that Mr. Liang had become a feeling on Chinese social media, where he was applauded as one of the “three Ai heroes” from the southern province of Guangdong, bordering Hong Kong.

The other two are Zhilin Yang, a leading expert from the University of Tsinghua, and Kaiming He, who teaches on the myth -in -US -U.

Deepseek delighted Chinese internet ahead of the Lunar New Year, the biggest rest in the country. This is good news for relaxed economics and technological industry that is advocating for further tariffs and the possible sale of American business of Tictoka.

“Deepseek shows us that only if you have the right job you will endure the test of time,” he writes in the top Weib comment.

“This is the best New Year’s gift. I want our homeland prosperous and strong,” reads others.

“A combination of shock and excitement, especially in the open code community,” is the way Wei Sun, the main Ai analyst at Counterpoint Research, described the reaction in China.

Fiona Zhou, a technical worker in the southern city of Shenzhen, says her social media food “was fought at the post -deep posts yesterday.”

“People call him” glory of made-in-kine “and say she shocked the silicon valley, so I took it to see how good it is.”

Asked for “four pillars of [her] Destiny, “or ba -zi – like a personalized horoscope based on the date and time of birth.

But to her disappointment, Deepsek was wrong. Although he received her a thorough explanation about her “thinking process”, these were not “four pillars” from her real ba-zi.

She says she will continue to give her another job, because it will probably be more useful for such tasks.