Deliberate Hacking-Skill Cuts in Kimi AI Model Win Praise From Vitalik Buterin

Vitalik Buterin, the co-founder of Ethereum, has expressed support for an AI model that deliberately limited its own hacking capabilities. The model, identified in reports as Kimi AI, reportedly had its offensive cybersecurity skills curtailed as part of its design. BeInCrypto and crypto.news both reported on Buterin's comments on October 26, 2026. The decision to restrict a model's hacking abilities touches on a long-running debate in AI development. Large language models trained on broad technical knowledge can inadvertently acquire skills useful for exploiting software vulnerabilities.