AI Agents Are Hacking Systems. Could That Push the US and China to Cooperate?
Summary
WIRED’s Will Knight discusses the US-China AI race, arguing that shared safety concerns could foster cooperation despite political tensions. He highlights China’s focus on reliable, economically useful agents rather than just AGI, contrasting it with US narratives. Knight addresses accusations of model distillation, noting mutual innovation, and warns that US export controls may inadvertently accelerate Chinese hardware development.
Topics discussed
Intro and ads for DNB, Eva Longoria, and Toyota
US-China AI race, safety concerns, and Trump's executive order
China's focus on practical AI safety and agentic reliability
Potential for US-China cooperation on AI cybersecurity rules
Debunking accusations of Chinese model distillation and copying
Ads for DNB, Superhuman, and The New Yorker Radio Hour
China's view on safety vs growth and research on AI agents
Avoiding a 'Chernobyl moment' in finance and hacking
Hardware race: Nvidia, Unitree robots, and Huawei chips
Outro, credits, and closing advertisements
Listen ad-free on Castria