InfoQ
1 sourcesZhipu's Tang Jie and GLM Team Disclose Early RSI Progress: GLM-5.3-Powered Infra Agent Optimizes Inference System
Zhipu founder Tang Jie and the GLM team published a long article disclosing that a GLM-5.3-powered Infra Agent completed model adaptation, system diagnosis, and performance optimization on a cluster of over 100,000 domestic chips, tripling end-to-end throughput from the initial baseline in under two weeks. The team believes this is not yet full recursive self-improvement (RSI) but an early form, as AI moves from helping humans develop models to participating in building its own successor.