White House AI Adviser Sacks Warns 'Cyber Guardrails' Are Handicapping U.S. Models Against China's Kimi K3

A senior Trump administration technology adviser has publicly argued that the safety restrictions built into American AI coding models are undercutting U.S. competitiveness, citing a case in which Moonshot AI's Kimi K3 fixed security bugs that rival U.S. systems declined to touch. The intervention thrusts a long-running debate among security researchers into the political arena, pitting deliberate safety design against practical performance in defensive cybersecurity work.