Garry Tan wants US open-weight AI labs to 'distill' frontier models, too
Points and comments are a snapshot, not live.
YC CEO Garry Tan wants US open-weight AI labs to also distill frontier models, opposing regulation.
Y Combinator CEO Garry Tan told CNBC he opposes regulating Chinese AI labs' distillation of US frontier models, and instead argues US open-weight labs should use the same techniques. He says proprietary labs like Anthropic overreach by controlling what customers can do with API outputs, and notes they trained on copyrighted data without permission. Tan warns that a single monolithic AI company would be the true doomer scenario.
What commenters are saying
Commenters broadly agree with Tan, arguing frontier labs have no moral high ground after training on copyrighted data without permission. Several point out OpenAI and Anthropic's terms of service assign output ownership to customers, undermining claims of illicit distillation. Some note YC's incentive to benefit portfolio companies via open-weight models. Others criticize the "lab" branding as misleading marketing from what are essentially AI companies.