Y Combinator CEO Garry Tan is pushing back hard against calls for strict regulatory crackdowns on AI model distillation. In a recent interview with CNBC, he argued that U.S. regulators should stay out of the practice entirely. In fact, he suggested American open-weight labs should actively adopt these techniques to build a robust domestic alternative to Chinese competitors. “We could argue that there should be an American distillation regime,” Tan stated, framing the issue as a strategic necessity rather than a legal violation.
This stance directly contradicts recent alarms raised by frontier model developers. Anthropic recently released a report alleging that Chinese labs are conducting “illicit distillation attacks” using stolen credentials and fraud. Anthropic CEO Dario Amodei has publicly urged regulators to intervene. Tan, however, sees this as corporate overreach. He argues that proprietary labs have no moral high ground to dictate how customers use API outputs, especially given their own history of scraping vast amounts of copyrighted human knowledge without permission to train their initial models.
Tan’s argument rests on a specific distinction: he isn’t advocating for fraud or stolen access. Instead, he wants smaller American labs to be free to “come in the front door” and learn from larger models through legitimate prompting. He told TechCrunch that controlling user behavior via restrictive terms of service feels constraining. He believes intelligence trained on broad public data should function more like a public good. This perspective challenges the current industry norm where closed-weight models tightly guard their reasoning processes.
The underlying fear here isn’t just about competition; it’s about monopoly. Tan warns that the true “doomer scenario” for artificial intelligence is not rogue agents, but market consolidation. If only one company controls the best capital and researchers, it becomes a monolithic power. By encouraging distillation, Tan hopes to keep the ecosystem fragmented and competitive. It is a bold bet that open-weight freedom will prevent any single entity from running away with the entire industry. Whether regulators will agree remains to be seen.
