Expert claims Grok 3 highlights AI divide between China and U.S.
Grok 3 highlights the disparities in artificial intelligence between China and the United States, according to an expert.
TroibNews 
The launch of Elon Musk's Grok 3 AI chatbot has sparked renewed discussions regarding the global AI competition. However, according to Tian Feng, former dean of the Intelligence Industry Research Institute at the Chinese AI software giant SenseTime, its importance lies not in its technical advances but rather in the differing development strategies of China and the U.S.
The U.S. approach is characterized by substantial funding and significant computing power to propel AI forward, while China opts for a more budget-conscious and efficiency-oriented method, exemplified by DeepSeek.
In a phone interview with CN, Tian explained that U.S. companies, supported by vast capital and advanced chips, drive AI advancements primarily through immense computational strength. This strategy, seen in the development of Grok 3, emphasizes innovation without regard for costs. With its financial dominance, the U.S. prioritizes maintaining a technological advantage over China rather than focusing on cost efficiency.
On the other hand, China's strategy revolves around optimizing the use of existing resources. DeepSeek's highly refined training and inference processes serve as a case in point. DeepSeek V3's training expenses are roughly one-tenth that of OpenAI's GPT-4, and experts predict that AI computing costs will further decline significantly over the next year. This efficiency-centric model promotes accessibility and accelerates the adoption of AI as a practical resource across various sectors.
The expert highlights the critical role of open-source development in the advancement of AI. Unlike Grok 3, which is proprietary, DeepSeek openly shares its research and permits commercial use without licensing fees. This transparency attracts a global pool of developers, creating a "flywheel effect" that enhances participation, leading to faster iterations, commercialization, and scientific breakthroughs.
Tian argues that for China to develop a robust AI ecosystem, it needs to foster a strong AI community through open collaboration. Open-source initiatives will help unite Chinese researchers and developers, promoting innovation on a broader scale.
Beyond AI models, Tian identifies China's manufacturing industry as a significant opportunity for AI integration. He notes that Tesla's gigafactories provide Elon Musk with abundant real-world data, giving him an advantage in AI development.
However, Tian believes that China's industrial diversity could be up to 100 times greater than that of the U.S., along with a larger scale.
The integration of AI into manufacturing could be transformative for China. As Tian states, "No scale, no intelligence." The progression of AI, particularly towards Artificial General Intelligence, relies heavily on extensive data and widespread adoption. With its vast industrial base, China is well-positioned to spearhead AI-driven advancements in manufacturing.
Looking ahead, Tian foresees two main trajectories for AI development:
1. High-cost, high-performance models – The U.S. will persist in heavily investing in elite AI models aimed at cutting-edge applications, utilizing its financial capabilities to further widen the gap with competitors.
2. Affordable, mass-scale AI – China will prioritize making AI more accessible and efficient, maximizing performance via current hardware and cutting costs.
As the global AI competition heats up, DeepSeek has emerged as a benchmark, with major models from both the U.S. and China being evaluated against it. Tian believes that DeepSeek's efficiencies and cost advantages will help it remain competitive in the changing AI landscape.
Ultimately, the future of AI will be influenced by these diverging approaches—one fueled by unlimited resources, the other focused on maximizing efficiency. The pressing question remains: which strategy will prove more sustainable in the long term?
Jessica Kline for TROIB News
The U.S. approach is characterized by substantial funding and significant computing power to propel AI forward, while China opts for a more budget-conscious and efficiency-oriented method, exemplified by DeepSeek.
In a phone interview with CN, Tian explained that U.S. companies, supported by vast capital and advanced chips, drive AI advancements primarily through immense computational strength. This strategy, seen in the development of Grok 3, emphasizes innovation without regard for costs. With its financial dominance, the U.S. prioritizes maintaining a technological advantage over China rather than focusing on cost efficiency.
On the other hand, China's strategy revolves around optimizing the use of existing resources. DeepSeek's highly refined training and inference processes serve as a case in point. DeepSeek V3's training expenses are roughly one-tenth that of OpenAI's GPT-4, and experts predict that AI computing costs will further decline significantly over the next year. This efficiency-centric model promotes accessibility and accelerates the adoption of AI as a practical resource across various sectors.
The expert highlights the critical role of open-source development in the advancement of AI. Unlike Grok 3, which is proprietary, DeepSeek openly shares its research and permits commercial use without licensing fees. This transparency attracts a global pool of developers, creating a "flywheel effect" that enhances participation, leading to faster iterations, commercialization, and scientific breakthroughs.
Tian argues that for China to develop a robust AI ecosystem, it needs to foster a strong AI community through open collaboration. Open-source initiatives will help unite Chinese researchers and developers, promoting innovation on a broader scale.
Beyond AI models, Tian identifies China's manufacturing industry as a significant opportunity for AI integration. He notes that Tesla's gigafactories provide Elon Musk with abundant real-world data, giving him an advantage in AI development.
However, Tian believes that China's industrial diversity could be up to 100 times greater than that of the U.S., along with a larger scale.
The integration of AI into manufacturing could be transformative for China. As Tian states, "No scale, no intelligence." The progression of AI, particularly towards Artificial General Intelligence, relies heavily on extensive data and widespread adoption. With its vast industrial base, China is well-positioned to spearhead AI-driven advancements in manufacturing.
Looking ahead, Tian foresees two main trajectories for AI development:
1. High-cost, high-performance models – The U.S. will persist in heavily investing in elite AI models aimed at cutting-edge applications, utilizing its financial capabilities to further widen the gap with competitors.
2. Affordable, mass-scale AI – China will prioritize making AI more accessible and efficient, maximizing performance via current hardware and cutting costs.
As the global AI competition heats up, DeepSeek has emerged as a benchmark, with major models from both the U.S. and China being evaluated against it. Tian believes that DeepSeek's efficiencies and cost advantages will help it remain competitive in the changing AI landscape.
Ultimately, the future of AI will be influenced by these diverging approaches—one fueled by unlimited resources, the other focused on maximizing efficiency. The pressing question remains: which strategy will prove more sustainable in the long term?
Jessica Kline for TROIB News
