Nathan Lambert, head of the post-training team at the Allen Institute for AI, has cautioned that U.S. legislative efforts to ban model distillation could negatively impact American open-source AI startups and academic research. Lambert argues that while distillation may erode the competitive edge of major companies like OpenAI, a ban could hinder innovation and concentrate power in closed-source labs. Lambert highlights that U.S. startups and academia rely on open-source models, often from China, to remain competitive. He warns that a ban could delay open-source advancements by six to twelve months and diminish U.S. influence in the global open-source community. Lambert emphasizes the importance of affordable open-source models in balancing the dominance of closed-source alternatives, which drive investment and innovation in the U.S. market.