Ethereum Founder Backs Moonshot's Safety-First Kimi 3.1 Model, Urging Capability-Based AI Restrictions Over Identity Gatekeeping

Vitalik Buterin has publicly endorsed Moonshot AI's decision to deliberately exclude offensive cybersecurity capabilities from its new Kimi K3.1 Pro model, released October 24, 2026, arguing that restricting AI by use case rather than user identity is essential to prevent dystopian regulatory overreach. The 3-trillion-parameter model, which features a 1-million-token context window, represents Moonshot's most capable open-weight foundation model to date and includes notably enhanced coding, financial research, and 3D modeling capabilities. In a detailed technical post, Buterin stated that limiting AI capabilities "by what you can do" rather than "by who you are" represents a far preferable path compared to identity-based access controls.