What you'll learn
Solve real problems with simple harnesses
See how a minimal agent loop and a small toolset can deliver useful outcomes without copying every frontier feature.
Avoid agentic complexity traps
Decide which harness capabilities your users need—and which create brittleness, cost, and maintenance work.
Build less and get better outcomes
Start with user needs and verification, then add complexity only when evidence shows the simpler system falls short.
Why this topic matters
AI FOMO makes every new harness feature feel mandatory. You ain’t gonna need most of it. Hugo will show how simple agent harnesses solve real user problems, why copying Claude Code is usually the wrong move, and how to choose the smallest set of capabilities your users need. Successful AI isn’t about keeping up. It’s about solving real problems.
You'll learn from
Coached teams at
