In the span of one week in late July 2026, three open letters from different factions of the AI industry laid bare a fundamental disagreement about how to govern frontier model development. The letters address open weights, distillation, and whether the pace of AI progress should be deliberately slowed.
Letter 1: Microsoft's open weights coalition (July 24)
Microsoft organized a letter titled "Open Weights and American AI Leadership" signed by 235 companies including NVIDIA, Amazon, Y Combinator, The Linux Foundation, and OpenAI. The letter argues that relying solely on closed models creates single points of failure and that open weight models allow researchers to examine behavior, identify vulnerabilities, and develop safeguards.
The letter takes a notable position on distillation, the practice of training one model on the output of another. It argues that distillation is a legitimate technique reflecting a long tradition of building on existing technology, and that policymakers should not conflate it with misappropriation. This position directly contradicts Anthropic's stance.
Letter 2: Anthropic breaks ranks (July 27)
Anthropic, notably absent from the Microsoft letter, published its own response three days later. CEO Dario Amodei stated that Anthropic has never advocated for a ban on open-weights models, but outlined two concerns that keep him up at night.
His primary concern is that authoritarian governments will build AI models more powerful than those built by the US and use them for military superiority or domestic repression. His secondary concern is that open-weights models present a higher risk for cyberattacks and biological attacks because guardrails cannot be applied or monitored once weights are released.
Amodei called for a crackdown on industrial-scale distillation operations, putting him directly at odds with the Microsoft letter's defense of the practice. He also reiterated support for chip export controls to China, rather than restrictions on open-weights software.
Letter 3: 1,337 frontier employees ask for pacing (July 28)
The third letter, "Pacing the Frontier," was signed by 1,337 employees of frontier AI companies. Signatories include Jakub Pachocki (Chief Scientist, OpenAI), Ilya Sutskever (Safe Superintelligence Inc), Dario Amodei and Jack Clark (Anthropic), and John Schulman. The letter asks the US government to support an international effort to develop tools for deliberately pacing the frontier of automated AI development.
The concern is competitive pressure combined with accelerated progress from automated AI research. Anthropic reportedly produces 80% of its code with Claude Code. OpenAI's Sol reduced serving costs by 20% through AI optimization. Kimi K3 designed a chip for its own architecture. The letter argues that without coordination tools, no single company or country can afford to slow down unilaterally.
The distillation fault line
The most concrete disagreement across the letters is on distillation. Microsoft's letter defends it as a standard technique. Anthropic has published research on detecting and preventing distillation attacks and explicitly called for enforcement against industrial-scale operations. This is not abstract: Anthropic recently accused Alibaba of running the largest Claude distillation campaign documented to date.
Open weight advocates argue that distillation is how the field advances and that restricting it would entrench incumbents. Critics argue that unauthorized distillation of proprietary models undermines the business model that funds safety research.
What the letters do not say
None of the letters proposes specific regulatory mechanisms. Microsoft's letter does not define what counts as "legitimate" distillation versus misappropriation. Anthropic's letter does not specify what a "crackdown" would look like. The pacing letter does not name a desired pace or a enforcement mechanism. The disagreement is real, but so is the gap between stating concerns and designing policy.
Sources
Open Weights and American AI Leadership - Microsoft
