AI Engineering Blog
Articles on AI engineering, LLMs, and building production-ready applications.
-
Compacting Conversation History: The Basics
How to keep a chatbot's conversation history from overflowing the context window by summarizing it, and why getting the summarizer prompt right is trickier than it sounds.
-
Stop Assuming Your AI Thinks Like You Do
Why LLMs fail at math, logic, and analysis - and how understanding them as statistical next-word predictors helps you use them effectively.
-
Skills Vs. Tools
An analysis of using agent skills versus agent tools.
-
Incorporating Agent Skills into Your LLM-Powered App: An Introduction
An introduction to using agent skills within an llm-powered app.
-
Building a Coding Agent: Part #1
How to build a coding agent from scratch.
-
Handing Your Filesystem to an Agent Using a Local Shell
How to let your LLM run shell commands on your filesystem using OpenAI's local shell.
-
Using a Hosted Shell for LLM Code Execution
How to let your LLM run arbitrary code safely using OpenAI's hosted shell, with a simple math-agent example.