Post

Calls to slow down AI development are about market position, not safety. Open-weight models already rival proprietary tech in many engineering and automation tasks. Lowering inference costs threatens closed-source business models, driving those companies to push for pre-deployment licensing. (1/2)

14/9/26, 6:36 am1 repost2 likes

Erik Hellman

@hellman.io

14 Sept

Licensing creates a classic regulatory moat. US AI companies can easily absorb compliance costs, while open-weight and international competitors (like Kimi or GLM) face massive friction. It’s corporate self-preservation disguised as safety to consolidate US control. (2/2)