The AI Giants Unite: Nvidia, Microsoft, Meta Lead Open Letter on Open Weights
On Friday, 25 companies and groups, including Nvidia, Microsoft, Meta, and others, published an open letter titled "Open Weights and American AI Leadership," urging Washington to support open-weight AI.
Jensen Huang, CEO of Nvidia, initiated the letter with his first-ever tweet, emphasizing the importance of both frontier closed models and open models for AI’s future.
Key Points:
- The industry is dividing into two camps: those supporting open weights (Nvidia, Microsoft, Meta, etc.) and those advocating for closed models (OpenAI, Anthropic, Google).
- The open letter highlights the benefits of open-source software in the 1980s and applies a similar argument to AI, stating that America’s AI leadership will be judged by its ability to build a strong, open ecosystem.
- Signatories include major tech companies like Mistral, Palantir, IBM, Andreessen Horowitz, Hugging Face, and Mozilla.
- The absence of OpenAI, Anthropic, and Google from the list is notable, as these companies are leaders in closed-model AI.
- The letter addresses the "China problem," mentioning allegations that Moonshot distilled Anthropic’s Fable 5 to create Kimi K3, leading to discussions about AI safety and potential sanctions.
AI Safety and Open Weights:
The open letter presents a compelling argument for open weights, stating that relying solely on closed models poses security risks as they can be "breached, misused, or fail in ways that outsiders cannot detect." In contrast, openness is presented as a path to enhancing AI safety. This direct response to OpenAI and Anthropic’s security-focused arguments could shape the ongoing debate about Chinese open-model makers.