Zhao Hejuan, the founder, Chairperson, and CEO of TMTPost, describes the actual gap between China and the U.S. in AI development as being significantly larger than what is commonly perceived. She believes that the gap is not just one to two years, as many argue, but rather it is a generational innovation gap. She emphasizes that the rapid development of models similar to GPT-3 and GPT-4 in China does not necessarily indicate that the gap is only one to two years. Instead, she suggests that this rapid development is primarily due to open source practices, and the underlying gap in innovation and capabilities remains substantial. Zhao also highlights that the gap is not only in computing power but also in the innovation capabilities of the underlying models and data capabilities.
How does Zhao Hejuan describe the actual gap between China and the U.S. in AI development, particularly in relation to generational innovation cycles?
Sources:
China’s DeepSeek Quietly Releases New Domain-Specific AI Model
China Shifts Used Cooking Oil Exports to Europe as U.S. Tariffs Bite
Volkswagen Warns of Profit Pressure From EV Transition, Trade Tensions
Stellantis Halts 2025 Forecasts Amid Uncertainty Over Trump Tariffs
Mercedes Withdraws 2025 Outlook Citing Trump Tariff Uncertainty, Profit Slumps in Q1
China’s Export Orders Slide in April as Trump Tariffs Take Toll on Manufacturing Sector
CATL Set to Launch $5 Billion Hong Kong Listing Next Week
China Drafts 'Whitelist' of U.S. Goods Exempt from 125% Tariffs, Sources Say
China Set to Quadruple GM Corn Planting in 2025 as Biotech Push Gains Pace
Meta Launches Llama API to Court AI Developers and Challenge Rivals