Garry Tan backs open US distillation arms race against Chinese AI labs

Y Combinator CEO Garry Tan is pushing back against Anthropic's alarm over Chinese AI labs using distillation techniques to extract knowledge from frontier models. Rather than tighten rules, Tan told CNBC and TechCrunch this week that the U.S. should embrace what he calls an "American distillation regime"—allowing smaller domestic open-weight labs to learn from proprietary frontier models legitimately. Tan rejects the premise that frontier AI labs should control how users and customers employ their models. He argues that frontier labs themselves freely ingested copyrighted material and broad public data without permission, so restricting downstream distillation feels hypocritical. His broader concern: if regulatory crackdowns consolidate power among a handful of proprietary providers, the real AI risk emerges—a monolithic landscape where innovation and access concentrate in one company.