Lesson 1.6 · 3:47
Same brain, different harness
Goal: Know what changes when you swap the model, and what changes when you swap the harness.
Try it
Sort each name into brain (model) or body (harness). If a name could be both, say which part you mean.
Claude · Claude Code · OpenCode · Codex · Grok · DeepSeek · GPT
Then, optionally, ask any chatbot you already use: Is OpenCode a model or a harness? Explain in one sentence. Check its answer against this lesson.
You're done when every name is sorted, and you've written one "thinking problem vs body problem" example of your own.
(Suggested: brain = Claude, Grok, DeepSeek, GPT. Body = Claude Code, OpenCode, Codex. Grok is also a harness name when you mean its command-line tool.)
Recap
- Swap the model → different thinking. Swap the harness → different context, tools, permissions and loop.
- Names overlap, so ask "brain or body?"
- Learn the pattern once. We use OpenCode for everything from here.
Next lesson (2.1): Module 2 begins. Files and folders, the agent's workspace.