Harrison Chase¶
Co-founder and CEO of langchain. Previously led ML at Robust Intelligence and the entity-linking team at Kensho. Stats + CS at Harvard.
At the Coding Agents Conference 2026 (with sam-partee), presented the thesis coding agents are the foundation for general-purpose agents — see harrison-sam-general-purpose-agents-2026. Core artifact: LangChain's Deep Agents harness, with a built-in planning tool, file-system-backed memory, and sub-agent delegation.
Grounded quotes¶
- "Things that we use every single day, right? Claude code… What does it have access to? A file system. It reads and writes files. It is able to over time, over its iterations, keep an idea of what it's trying to do."
- On memory: a retrievable object outside the context window — "extend not only through just simple search like rag, but all the way through tool calling."
- On the agent identity shift: "with Open Claw, I think we saw actually these agents take on their own identity, which I think was really interesting… it has its memory, and it has its credentials."