About

Agentic Systems is a blog about building AI agents: software in which a language model chooses its own next action, uses tools, and works toward a goal over many steps.

How we write

• One idea per post. Each post covers one idea that holds up over time, such as the agent loop, context engineering, multi-agent topologies or agent evals. We explain how it works and trace where it came from.
• Our own words, our own diagrams. We learn from papers and engineering write-ups, and link to every source. We redraw every diagram and write every explanation ourselves.
• Code you can run. Most posts link to a small, framework-free example in our examples repo.
• Trade-offs included. Each post says when not to use the idea.