Open-source software powered the internet era, and open-weight AI models are poised to do the same for the AI age. Despite concerns from some US officials, exemplified by Treasury secretary Scott Bessent’s suggestion of sanctions, restricting these publicly released and customizable models would be misguided. China's advancements, such as Moonshot's Kimi K3 model, demonstrate their progress even with limited chip access, potentially through techniques like distillation.

This has led to investor worries about the profitability of large US AI labs that rely on closed models, particularly as cheaper Chinese alternatives emerge. While legitimate concerns about intellectual property and safety, including the potential misuse of open-weight models for harmful purposes, need addressing, protectionism is not the answer. An open letter signed by leaders from Nvidia, Meta, and OpenAI emphasizes the benefits of an open ecosystem for innovation and competition.

Companies like Anthropic, represented by Dario Amodei, advocate for mandatory safety testing rather than outright bans. The US should foster its own open-weight models to compete and encourage international cooperation on standards, rather than restricting China's contributions. This approach can drive AI adoption across the economy, ultimately benefiting the US tech industry through increased prosperity and growth.