Why Your AI Agents Don’t Work with Dex Horthy of HumanLayer | In-Context Cooking
Source
youtube.com
Author
Latent Space
Date
Context
A formal research-and-planning process can waste time on a tiny code change and help with a task an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → cannot solve directly. In this interview, the guest describes learning when to use a research, plan and implement workflow by trying both approaches. A one-line edit may need only a direct request; a larger problem may leave the agent flailing because it lacks well-chosen context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition →. The guest's advice is to build judgment from those failures and successes. The amount of preparation should follow the difficulty the agent actually encounters, rather than become a fixed ritual for every coding request.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
context engineering — Deciding what an agent sees at each step — the discipline that replaced prompt engineering once context windows got large enough to fill badly.
