The sudden arrival of the Chinese artificial intelligence startup Deepseek breakthrough AI model has caused the Silicon Valley elite’s anger, jealousy and admiration. It shows the ability to compete with AI leaders such as Openai, Meta (META), and humans. The model of Deepseek is a small part of its American peer cost. This strong move surprised many technology industries.
DeepSeek’s V3 model was released in December, and then the new reasoning model earlier this month. Only 2,000 NVIDIA (NVDA) expensive graphics processing unit (GPU) was created. This unit (GPU) was based on Hangzhou -based startups. Essence Last year, hedging fund climax. The company’s self -proclaimed capabilities of the American AI giant’s betting on the American AI giant’s betting. NVIDIA’s stock price fell 17 % today (January 27). AMD, Alphabet (GOOGL) and Microsoft (MSFT) have also declined.
Just as some technology leaders question whether Deepseek has more AI chips than allowed, others advocate stricter GPU export control measures to prevent China from gaining advantages in new technologies. Risk capitalist Marc Andreessen, known as Deepseek’s breakthrough, is a modern “artificial satellite moment”. Satya Nadella, Microsoft, praised the company’s efficiency, and Meta’s chief AI scientist, Yann Lecun, praised his decision to decide open source.
It is certain that the development of DeepSeek has led Silicon Valley. The following is what the famous technical figure says the surprising success of this startup:
Marc Andreessen, co -founder of Andreessen Horowitz
The ordinary partner of the venture capital company Andreessen Horowitz, Marc Andreessen, described the “artificial satellite time” called AI in the X post of AI yesterday (January 26) in the R1 modification model released on January (January 26), which was released on January 20 (January 26). Point out how unexpected success. The satellite launch of the Soviet Union in 1957 stimulated the United States’ strengthening efforts. DeepSeek-R1 “is one of the most amazing and impressive breakthroughs I have ever seen,” Andreessen added that he praised the company’s decision to open the model as “profound gift for the world.”
Meta chief AI scientist Yann Lecun
According to Yann Lecun, Deepseek’s performance does not necessarily show that China’s AI capability exceeds the US function, but emphasizes the power of open source. “The correct reading is: ‘The open source model exceeds the proprietary model,” said computer scientists in an article about Meta threads recently.
Lecun also said that AI stocks were sold to respond to Deepseek “serious unreasonable” and pointed out that many monetary technology companies spent funds on AI that they were not necessarily used for training models, but they were running them.
Human CEO Dario Amodei
For some technical leaders, the breakthrough of Deepseek indicates that the United States should continue to fight chip export control. Dario Amodei recently told CNBC that he believes that China’s artificial intelligence company’s GPU has exceeded expectations because many people can purchase a large amount of hardware before implementing such restrictions. As far as DeepSeek is concerned, “at least it is reported that they have 50,000 groups.”
According to the anthropomorphic CEO, export control must be concentrated in preventing tens of thousands of chip inventory. “I think if the United States cannot lead this technology, we will be in a very bad position in geopolitics,” Amodei told CNBC.
Salesforce CEO Marc Benioff
For Marc Benioff, the popularity of the new discovery of DeepSeek proves that data (rather than calculation or model) will make the technical company’s progress. “The real treasure of AI is not [user interface] Benioff said in an article on X yesterday that Benioff said on X that the success of Deepseek was obtained without a huge fund or NVIDIA chip. “The real value lies in data and metadata, and the potential of oxygen is AI. “
Microsoft CEO Satya Nadella
Saty Nadella praised the progress of this Chinese startup when talking with CNBC in Davos. “I think we should attach great importance to China’s development.”
CEO published his comments today. Among them, X posts seemed to associate Deepseek with “Jevons Paradox”. Jevons Paradox improved efficiency and caused resource consumption rather than decrease. He said: “As artificial intelligence becomes more efficient and easy to use, we will see its soaring use and turn it into a product we can’t get at all.”
