Quote · The Prof G Pod with Scott Galloway
China Decode: Hong Kong's AI Crackdown, Lululemon’s Marketing Backlash, and World Cup Fever
Where this was said
Are U.S. AI Restrictions Actually Working?
At 17:09 · chapter starts 16:40
James Kynge delivers a blunt verdict: U.S. technology restrictions on China have not worked when it comes to AI. Banning Hong Kong while leaving Singapore, the UK, and Silicon Valley fully open to Chinese engineers defeats the purpose entirely. He estimates China is roughly a few months behind the U.S. in developing world-class LLMs — a gap that restrictions have done little to widen [1] — James Kynge "A Chinese engineer in Singapore, or even in Silicon Valley, can still access Anthropic and train Chinese LLMs on U.S. models. Banning Hong …" 16:49 . The hosts note that David Sachs, one of the few voices in the White House arguing export restrictions were counterproductive, has now left his role, which Alice Han fears will only accelerate a more aggressive and ultimately ineffective crackdown.
A Chinese engineer in Singapore, or even in Silicon Valley, can still access Anthropic and train Chinese LLMs on U.S. models. Banning Hong Kong is symbolic theater, not effective policy. The U.S. has failed to meaningfully slow China's AI development.
Alice Han noted that financial services account for roughly a quarter of Hong Kong's GDP and employ 250,000 people, making AI access restrictions particularly damaging.
Hong Kong once promised R&D teams access to both U.S. and Chinese frontier AI models — the best of both worlds. That vision is dead. With U.S. restrictions now lumping Hong Kong with the mainland, the city's unique cross-border value proposition is evaporating fast.
