Kyle Chan, an expert on China and AI, discusses the fundamental differences between U.S. and Chinese approaches to artificial intelligence development with the host of Interesting Times.

The conversation explores whether the two superpowers are truly in an 'AI race' and what winning even means in this context.

Chan argues that while the U.S. focuses intensely on achieving artificial general intelligence (AGI), China pursues multiple parallel strategies emphasizing efficiency, practical applications, and widespread deployment.

The discussion covers everything from chip export controls and energy infrastructure to public attitudes about AI and the possibility of future AI arms control agreements between the superpowers.

China's Multi-Track AI Strategy vs America's AGI Focus

The U.S. concentrates on AGI development with massive investments in companies like OpenAI and Anthropic, now valued close to trillion dollars each, betting on achieving artificial superintelligence.

China runs multiple parallel races: improving AI models to keep pace, developing efficiency to make models smaller and cheaper, pursuing diffusion through open source, and focusing heavily on robotics applications.

"You don't really hear so much about AGI" in China - Kyle, noting that Chinese tech founders sometimes echo U.S. rhetoric but prioritize practical applications in daily life.

Everyday AI: Robots Serving Food While Americans Debate Safety

In major Chinese cities, visitors would notice autonomous delivery robots handling packages and food, robot waiters in restaurants, and hotel room service delivered by robots rather than human staff.

Self-driving cars and drone delivery for coffee represent subtle but surprising differences from typical American AI interactions, which remain largely digital.

This physical-world AI deployment reflects China's focus on practical applications over the theoretical superintelligence goals driving Silicon Valley.

The Chip Wars: Export Controls and Energy Advantages

U.S. export controls prevent China from accessing the most advanced NVIDIA chips, forcing them to use watered-down versions while America keeps the best semiconductors for domestic companies and allies.

NVIDIA doesn't manufacture chips - Taiwan's TSMC produces them using Dutch company ASML's lithography machines, creating a complex global supply chain that U.S. controls effectively cut China out of.

China's major advantage is energy infrastructure: rapid buildout of solar, wind, and battery capacity that can power data centers, particularly in western provinces away from urban areas.

"For the US, this is a major bottleneck. It's very hard now for data centers to build out the power capacity" - Kyle, contrasting American energy constraints with Chinese abundance.

Opposite Anxieties: Chinese Fear of Falling Behind

Chinese public anxiety focuses on not keeping pace with AI technology and losing competitiveness in the labor market, contrasting sharply with American fears about job displacement and AI risks.

Youth unemployment in China reaches 17%, double the U.S. rate, with 12 million new college graduates annually competing for white-collar jobs and viewing AI skills as essential for employment.

"The fear among individuals and companies and workers is that they're not keeping pace with AI, that they're not using it enough" - Kyle, describing how individual anxiety mirrors national-level concerns.

This anxiety amplifies existing hyper-competitive pressures in Chinese society, where workers desperately want to avoid factory or delivery jobs in favor of prestigious positions.

Beijing's Social Engineering: Cracking Down on AI Girlfriends

Chinese regulators are already implementing policies against AI boyfriends and girlfriends, viewing them as non-productive time wasters similar to their previous crackdowns on video games and private tutoring.

"They have a very sort of negative view of wasting time, basically, of what they see, the folks in Beijing, what they see as sort of non-productive activity" - Kyle on government attitudes.

The crackdown reflects concerns that AI companions could become time sinks for Chinese youth who should be "engineering the future and building the startups and the future Chinese versions of SpaceX."

Simultaneously, robots are seen as solutions to China's shrinking workforce and aging population, with humanoid robots viewed as potential factory workers to maintain manufacturing competitiveness.

The AGI Question: Beijing Isn't Drinking the Kool-Aid

"I don't think Beijing is an AGI-pilled" - Kyle, suggesting Chinese leadership doesn't share Silicon Valley's transcendent view of artificial general intelligence.

China declined Trump's offer of relaxed export controls on H200 NVIDIA chips, prioritizing semiconductor independence over short-term AI gains, indicating they're not desperately sprinting toward AGI.

Chinese AI companies like DeepSeek may believe more strongly in superintelligence than Beijing policymakers, creating a gap between commercial ambitions and state priorities.

China's approach mirrors their historical integration of internet and IT infrastructure into basic services rather than pursuing revolutionary technological breakthroughs.

Espionage Through Distillation and Policy Implications

Chinese AI labs appear to be conducting unauthorized "distillation" - training weaker models on outputs from stronger American models like Claude - using proxy accounts to circumvent restrictions.

This technique resembles Microsoft's historical battles against black market Windows copies rather than traditional IP theft, representing a new category of technological competition.

Chan argues the U.S. should "take a step back from this all-out race framework" because the race mentality drives recklessness and eliminates safety guardrails unnecessarily.

The U.S. should focus more on deployment and open source models to compete with China's diffusion strategy, rather than just pursuing the highest-end capabilities.