In a recent statement by Alexandr Wang, the founder of the American artificial intelligence company Scale AI, he mentioned that the United States is ahead of Beijing in the AI competition, with China’s AI model essentially being a copycat of the American model.
Scale AI primarily operates in the data labeling business and has seen its market value grow from zero to around $14 billion over the span of eight years, making it one of the unicorn companies in the American AI industry. Scale AI has been closely associated with almost every major breakthrough in the field of AI. Additionally, it has collaborations with the U.S. Department of Defense and the global giant OpenAI.
“If you look at the best AI models in China, they are essentially replicas of the American model. This is why I believe that the default path is for the U.S. to be leading,” said Mr. Wang, the 27-year-old Chinese-American genius.
The field of AI is widely recognized to have three fundamental pillars – computing power, algorithms, and data. In terms of computing power, any large model cannot do without American chip company Nvidia, which also relies on Scale AI for providing high-quality data labeling services. Some liken Nvidia as the “shovel seller” of computing power and Scale AI as the “shovel seller” of data.
As the founder and CEO of Scale AI, Wang warned that the U.S. needs to monitor how China is leveraging the internet and overseas data. Currently, it seems that China’s model is essentially copying what the U.S. has.
Zhou Hongyi, the founder of the Chinese mainland cybersecurity company Qihoo 360, admitted in an interview with the CCP-controlled nationalist newspaper Global Times, that mainland China focuses on the application development in the AI field. However, it may take one or two years for Chinese AI companies to catch up once American AI companies open source their technology. Wang pointed out that while China has the concept of data factories, they exploit the data of billions of Chinese citizens for AI models’ services, such as facial recognition and autonomous driving. This gives China an advantage in AI development, but their performance has not met expectations.
“The current situation is even if China doubles its computing expenses, their efficiency does not match the efficiency of America doubling its computing expenses,” Mr. Wang remarked.
Japanese electronics engineer Li Jixin told Epoch Times that “even the most powerful AI needs massive data for training and support. Citizens in lawful countries often pay attention to protecting their privacy information, making it challenging for large tech companies to access vast amounts of personal data for AI training. In China, citizens cannot supervise the government, allowing the government to freely use private information of the people for big data training to gain an advantage in the AI field. For instance, China’s (CCP) surveillance technology and internet censorship technology far exceed those of other countries.”
To ensure that the U.S. maintains its lead over the CCP in AI technology, Mr. Wang mentioned observing the scaling rules of the Chinese model.
“We need to have a competitive intelligence team closely monitoring various model sizes and the computational loads entering these models to understand these scaling rules,” he suggested.
The “2023 Artificial Intelligence Index Report” by Stanford University noted that “artificial intelligence is expected to become a critical component of both economic and military power in the near future.” The development of artificial intelligence will reshape the geopolitical landscape.
To curb the development of CCP’s AI technology, the U.S. has implemented export controls on cutting-edge chips, limiting the CCP’s access to advanced computation chips, supercomputer development, and semiconductor manufacturing capabilities for military and other purposes. In retaliation, the CCP restricts the export of rare minerals.
According to Mr. Wang, there are three key indicators for measuring the AI ecosystem between China and the U.S.
The first indicator is the comparison of the quality of Chinese chips with American chips. The latest statistics show that in terms of computing power, the yield rate of Huawei’s Ascend 910B produced by China’s SMIC is only about 80% of Nvidia’s chips, but in terms of value for money, the price is actually two to three times that of Nvidia chips.
The second indicator is the production level. Huawei produces 100,000 chips every quarter, while Nvidia produces about 1 million chips each quarter. Wang warned that once this ratio reaches one to three, the U.S. needs to closely monitor and track it.
The third indicator is power supply. Wang mentioned that it’s crucial to closely measure the new power generation capacity between the U.S. and China because, in the mid-term, the real bottleneck for AI is the power supply, not the actual chips.
In the field of autonomous driving, Scale AI’s data engine has driven the breakthrough of L4 autonomous driving technology, providing power for many major AI projects of the U.S. Department of Defense through its public sector data engine.
Meanwhile, Scale AI has been actively collaborating with the U.S. public sector. In August 2023, the company signed a contract with the U.S. Department of Defense’s Chief Digital and Artificial Intelligence Office to help boost AI capabilities across the military, including the Army, Marine Corps, Navy, Air Force, Space Force, and Coast Guard.
In May 2024, Scale AI launched the AI-based decision platform Donovan, the first large language model deployed on the classified network of the U.S. government.