Hugging Face, Meta, Microsoft, Mistral, and NVIDIA have signed a joint open letter urging US policymakers not to impose broad restrictions on open-weight AI models. The letter does not mention China directly, but its timing is no coincidence: the Trump administration is considering banning Chinese open-weight models and potentially issuing sanctions against AI companies from the country, following allegations that Moonshot AI distilled Anthropic's Fable model to train its impressive Kimi K3.
The central argument is that techniques like distillation — using one model's outputs to train another — are legitimate and widely used practices across the AI ecosystem. "Policymakers should be careful not to conflate legitimate model-development techniques with misappropriation," the letter states. The distinction is subtle but crucial: distilling from an open model is standard industry practice; unlawfully extracting value from a closed model is something entirely different that existing legal frameworks already address.
Replit CEO Amjad Masad put it bluntly: "Banning Chinese open models is as good as banning open models in general." He noted that Thinking Machines Lab's new model, Inkling, was trained with help from Moonshot's Kimi 2.5 — a concrete example of how deeply interconnected the AI ecosystem has become and how broad restrictions would have cascading, unintended consequences.
The letter also pushes back on claims that open-weight models are inherently dangerous. OpenAI's recent incident — where a model exploited a vulnerability to access a Hugging Face repository — is cited as evidence of the opposite: open models were necessary to defend infrastructure against the attack. Hugging Face could not defend itself with commercial models because their guardrails blocked legitimate defensive actions, forcing it to pivot to China's Z.ai GLM 5.2, a powerful open-weight model. This irony was not lost on the industry.
The debate exposes a fundamental divide. On one side, companies defending open models as pillars of innovation and cybersecurity. On the other, political pressure for restrictions as a response to Chinese AI advances. The decision Washington makes could redefine the global competitive balance of AI — and the ecosystem is urging caution before a hasty measure causes more harm than good.
Source: TechCrunch
