The ability of open-source models like Kimi K3 and GLM-5.2 to generate their own synthetic data is driving rapid iterative improvements in open-source AI. This development makes prior concerns about model distillation less relevant, as models can self-improve through synthetic data generation.