TLDR; memory = harness => keep your memory and harness on your side of the API.

Harrison Chase (creator of LangChain) shared a concise architectural principle regarding AI agent design: memory and harness components should remain on the application side of the API boundary rather than being embedded within the model itself. This reflects a broader design philosophy in agentic systems where stateful components like conversation memory and execution harnesses are best managed by the orchestration layer. The post condenses a key LangChain/agent architecture guideline into a memorable shorthand for developers building LLM-powered applications.

your harness = your memory if you'd rather read in blog form, link here: https://blog.langchain.com/your-harness-your-memory/
memory lock-in doesn't kick in when you adopt the harness. it kicks in 6 months later when leaving means starting over from zero. by then th…
“Memory is important, and it creates lock in” Exactly why I don’t want OpenAI, Anthropic or any of the other AI companies owning it.
for months i’ve been giving talks telling people that owning their memory and skills folder is the most important store of value they curren…

Source: Harrison Chase (X) | 2026-04-12

Related