Garry Tan Pushes for US Open AI Models to Match China's Distillation Tactics
Y Combinator CEO Garry Tan has called for American open-weight AI labs to legally distill knowledge from frontier models.
This approach aims to create stronger domestic alternatives to Chinese AI systems.
Understanding Distillation in the AI Race
Distillation lets smaller labs learn advanced reasoning by carefully prompting top models through normal API access.
Frontier companies like those behind closed models have long trained on vast public data without always seeking permission first.
Tan argues that restricting what users do with model outputs could stifle healthy competition in the industry.
Such restrictions might limit innovation from smaller players who cannot afford to build everything from scratch.
Why Open Models Matter for Everyday Users
A balanced ecosystem prevents any single company from dominating AI power and access.
This matters to regular people because open models can lower costs for tools used in education, small businesses, and creative work.
Looking ahead, widespread distillation could speed up global AI progress while keeping options diverse.
History shows that open approaches have often driven broader adoption and unexpected breakthroughs in technology.
Overall, Tan's view highlights a push toward treating advanced intelligence as more of a shared resource than a locked service.
