LangChain's Harrison Chase on Goal Engineering: designing goals for goal-working agents

hwchase17 · x · 2026-11-17

Harrison Chase co-authored a new article, "Goal Engineering: How to Design Goals", with @mgarimella and Anshul Gupta of Actively AI, sharing lessons learned in production.

Core claim: the scope of tasks agents can take on has grown drastically, and as agents move toward working on goals, designing those goals correctly becomes critical.

Sydney Runkle adds: the most important type of context is domain knowledge and taste — often delivered as feedback, so review UX needs to be delightful for agents to continually learn, becoming more useful and more autonomous.

Related event: LangChain founder proposes Goal Engineering for the agent era(2 posts)→

Original post →

More from coding & agent

coding & agent channel →