Anti-distillation curbs could widen China's LLM gap with frontier labs to 12 months

An unnamed WeChat post titled "China LLM discussion notes," shared on X by @indigox, argues that anti-distillation measures are blocking China's main route to catch up with frontier AI labs.

The post says distillation costs have risen 5–10 times, and up to 20 times in some cases. It also says training 3–5 trillion-parameter models needs tens of thousands of B-class GPUs.

The post estimates the gap is 3–4 months with distillation and would grow to 12 months once anti-distillation takes effect. It further estimates 18–24 months if leading US labs reach recursive self-improvement next year. The source does not identify its author, and the report contains no independent evidence for these figures.