Posts

All the articles I've posted.

• Agentic AI & LLM Weekly — 2026-W09

  27 Feb, 2026

  This week exposed the fragility underneath frontier AI's polish: a single training prompt unaligns 15 models, LLM agents deanonymize you from your Reddit posts, and Anthropic reveals industrial-scale model theft from Chinese labs.