harness engineering > prompt engineering prompts get you 80% there the last 20% — tool routing, memory retrieval, retry logic, evals — thats…

Harrison Chase, co-founder of LangChain, argues that while prompt engineering is foundational and can get you 80% of the way to a functional AI system, the remaining 20% of engineering effort involves critical infrastructure concerns. These include tool routing, memory retrieval, retry logic, and evaluations (evals), which collectively determine whether an AI application is production-ready and reliable. This perspective highlights that building robust AI systems requires significant engineering work beyond prompt crafting alone.

here's how we're improving our base harness, you can apply these same lessons to hill-climbing for your application-specific harness!
Another banger article from the @LangChain team! Harness evolution combined with specialist local models will be the way forward undoubtedly…
your harness = your memory if you'd rather read in blog form, link here: https://blog.langchain.com/your-harness-your-memory/
The ‘agent harness’ is essentially the new runtime. It’s like a lightweight OS around an LLM—handling memory, tool calls, retries, and contr…
@hwchase17 Exactly. he harness is where the 'soul' of the agent lives. Excited to see how Deep Agents and LangChain can keep pushing this la…

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

Related