Tag
This blog post by SuJinyan6 examines the evolution of AI agents from simple LLM+tool use to context engineering and long-running harnesses, citing Anthropic's recent work and discussing how agent capability is now a system-level property involving multiple components.