Silicon Valley leaders and national security experts are concerned about Chinese companies using AI model distillation, a process of training smaller models on outputs from larger ones. Some U.S. agencies have issued advisories on the practice, with companies like Anthropic and OpenAI accusing Chinese firms of illicit distillation. However, Y Combinator CEO Garry Tan disagrees with calls for action against these companies. He believes regulators should focus on creating balance between open-weight and frontier AI models, ensuring frontier models remain commercially viable. Tan advocates for more open-weight AI options developed by American labs, using similar training techniques. He argues that proprietary AI labs also used vast amounts of public data, including copyrighted material, without explicit permission during their own training. Tan views controlling downstream use of AI outputs as an overreach by labs. He believes intelligence trained on broadly accessible data should be more of a public good. Tan's primary concern is preventing a single monolithic company from dominating the AI landscape. He suggests a regulatory approach promoting access and competition as the optimal outcome.