Garry Tan says “I would do nothing” about China's AI distillation and urges the industry to focus on current AI risks instead of doomsday-style extinction fears
At a time when some Silicon Valley giants and national security experts are calling for action against Chinese companies engaged …
Context & Ripple Effects
Tan’s intervention extends a long-running split over AI governance: in 2023, Demis Hassabis pushed back on accusations that prominent lab leaders were using catastrophic-risk rhetoric for regulatory capture. By February 2026, both Sam Altman’s concerns about adoption resistance and Jensen Huang’s warning about the “doomer narrative” had sharpened the argument over which risks deserve attention.
It also lands amid competing models of AI competition. Beijing has emphasized an application-oriented AI strategy, while Li Qiang paired a call for global AI cooperation with concern over chip-supply bottlenecks. Tan’s position treats diffusion of model capabilities and practical harms as more consequential than an extinction-centered agenda.
First-order effects
- Tan gives founders and open-weight AI advocates a high-profile argument against making enforcement over Chinese model distillation the industry’s central response.
- His emphasis redirects the policy debate toward immediate AI harms, challenging frontier-lab voices that foreground extinction risk.
Second-order effects
- Open-weight labs face greater pressure to explain whether broader model access and distillation advance competition without sidelining the practical risks Tan wants prioritized.
- Frontier labs and their policy allies must make the case that restrictions on model copying address a concrete competitive or security problem, rather than merely reinforcing a catastrophic-risk framing.
Third-order effects
- If application-led deployment continues to define China’s AI strategy while U.S. debate centers on frontier-model controls, AI governance may split more sharply between diffusion and adoption concerns on one side and capability containment on the other.
- The dispute points toward a durable contest over whether AI policy should be organized around control of leading models or the labor, market, and safety effects of their widespread use.
The trend: AI governance is increasingly dividing between frontier-capability containment and a more pragmatic agenda focused on model diffusion, adoption, and near-term harms.