In a Stratechery essay, Ben Thompson argues that concerns over Chinese open-weight AI models such as Kimi K3 are largely overstated from a US economic standpoint. He contends that frontier labs retain pricing advantages due to compute scarcity rather than any fundamental technical edge, even as inference costs fall industry-wide. Thompson calls for loosening restrictions on US open-weight models and on cybersecurity applications of frontier models, warning that current policy inadvertently pushes American companies toward Chinese models for defense-related uses.
