Everything Gets Rebuilt: The New AI Agent Stack | Harrison Chase, LangChain

The podcast explores the evolution and architecture of AI agents, focusing on the infrastructure required for agents that plan, use tools, and manage memory. Harrison Chase, co-founder and CEO of LangChain, discusses the importance of the "harness"—the framework that enables models to interact effectively with their environment—over the models themselves, citing Claude Code and Manus as examples where the harness was key to success. Chase differentiates between conversational agents, requiring low latency, and long-horizon agents, which often resemble coding agents due to code's versatility and models' training. He also highlights core components of modern agent architecture, including detailed system prompts, planning tools, sub-agents for context isolation, and file systems for LLMs to manage their context.