rotative
wired ai · July 24, 2026

OpenAI fears China’s Kimi K3, but VC Gurley wants it free

The White House accused China's Moonshot AI of distilling Anthropic models to build Kimi K3, sparking a debate where startups and VCs like Bill Gurley lobby against bans, warning that restrictions would only entrench US giants.

read in 简体中文 · 繁體中文 · 日本語 · 한국어

a fierce debate has erupted after the white house reported that beijing-based moonshot ai developed its kimi k3 model by distilling anthropic’s proprietary fable 5, following earlier accusations against alibaba. large labs like anthropic cite these incidents as reasons to regulate open-weight systems to protect ip and enforce safety guardrails. however, the 'little tech association' and investors like bill gurley and chamath palihapitiya are lobbying against restrictions, arguing that bans on foreign models create monopolies for us giants and deny capital-constrained startups access to affordable tools. the conflict highlights a growing split between frontier model providers protecting their margins and the broader ecosystem relying on open access.

read at source ↗

← rotative