What is Loop Engineering
"I don’t write prompts for Claude anymore. I write loops, and the loops do the work." -- Boris Cherny, Head of Claude Code, Anthropic
Loop Engineering is the approach where automated loops prompt the agents, not the human. Instead of typing prompts one by one, you build systems that launch, monitor, and iterate on AI agents in a continuous loop.
This course explores the approach without the hype. We cover the real advantages, hidden costs, technical implementation, and practical examples with prompts ready to copy. An honest take: loops are powerful, but they're not a silver bullet.
Course Tracks
Fundamentals of Loop Engineering
What it is, where it came from, and how the orchestrator-workers architecture works.
Advantages vs. Disadvantages
Critical analysis without hype. Real costs, hidden risks, and when NOT to use it.
Technical Implementation
Hands-on: Claude Code, Archon, control dashboard, and cost optimization.
Practical Examples and Prompts
Ready-to-use recipes: task runner, GitHub bot, multi-agent orchestrator, anti-patterns.
Who is behind it
Boris Cherny
Head of Claude Code, Anthropic
Manages tens of thousands of agents per day. Created the paradigm of loops replacing manual prompts in Claude Code.
Peter Steinberger
Creator of OpenClaw
Runs loops that prompt Claude on its behalf. A pioneer of the orchestrator + workers approach with external state management.
Cole Medin
Creator of Archon
Analyzes the approach in a critical, practical tone. His video is the main source for this course. He built Archon as a harness builder.
Ready to get started?
Start with Track 1 to understand the fundamentals. Each track builds on the previous one.