At a time when some Silicon Valley giants and national security experts are calling for action against Chinese model distillation companies, Garry Tan, CEO of the famous startup accelerator Y Combinator, is less worried.
“I wouldn’t do anything” about distillation, he told CNBC at Y Combinator’s annual Demo Day. “We could argue that there should be an American distillation regime.”
Distillation is the process of using the results of a more capable AI model to train a smaller or less capable one, sometimes illicitly. It has been a point of great concern for frontier model developers, namely OpenAI and Anthropic. Anthropic has accused Chinese companies such as Moonshot AI, DeepSeek and MiniMax of this practice, while OpenAI believes that the architectures of DeepSeek’s V3 and R1 models were distilled from its own GPT-4 and GPT-4o models.
Amid this, the National Security Agency, the US Cybersecurity and Infrastructure Security Agency and the Federal Bureau of Investigation on Tuesday released an official cybersecurity warning on the issue.
Critics have dismissed complaints about Anthropic and OpenAI’s distillation because much of the data used to train these AI models may be covered by copyright law, which Tan highlighted. That controversy is currently the subject of legal battles in the US, including the New York Times suing OpenAI and Microsoft in 2023 over unauthorized use of their papers in model training data, and a consortium of book authors settling with Anthropic in 2025 over a similar claim.
But Tan believes regulators should focus less on curbing distillation and more on creating a balance between open-weight models and frontier models, as long as frontier models maintain a price premium that allows their business model to remain viable.
“This is actually the ideal case. What you want is for open weight models to give people freedom and access,” he explained. “If I were a regulator, that’s what I would look for.”
Tan acknowledged that it is difficult to achieve a balance, calling it “a tightrope.” However, he says it’s worth seeking a balance: “it could lead to the best possible outcome.”
Tan also touts a more moderate approach to AI safety, where recent apocalyptic headlines surrounding the resignation of anthropic researcher Jacob Coxon have sparked record fears in the public about the existential risks of AI.
“We need to focus on scientific facts, not science fiction,” he said. “We need to respond to what’s happening right now. If there was a breach and a coordinated attempt by agents to take over our infrastructure, what do we do about it?”
Cybersecurity is a risk that Tan considers imminent. And there are other risks that are both existential and short-term. For example, on Thursday, Anthropic reported that it had blocked Claude’s access for five cases of scientists in unspecified foreign countries who were using the models to conduct research on dangerous pathogens. Anthropic feared that these researchers were covertly using Claude to create biological weapons.
Other risks, however, have a longer timeline. Tan believes that AI-related job losses and economic transformation will not have an immediate impact. Rather, it predicts that over time, people will move to automating routine tasks and devoting more work hours to creative activities.
“It will take decades for this to seep into society, and that’s not a bad thing,” he said.
That longer timeline hasn’t stopped Y Combinator from investing in AI right now. Of the 196 startups that presented at Demo Day, 149 were classified as machine learning and artificial intelligence companies.
Source link
