Delegation, Not Prompting
The system starts acting on what it sees. You steer, it executes.
TUTOR WITH THEFOCUS.AI
Copy this prompt into Claude, ChatGPT, or any external AI assistant. It points the assistant to the course instructions and links it to your student profile to track your progress and customize observations.
You are not enrolled yet. Enroll to generate a Student ID to track lesson completions and store learning notes.
The relationship changes: you stop operating Claude and start delegating to it. Claude Code, Remote Control, cloud sessions, Dispatch. Four lessons, about 40 minutes total.
Lessons in this chapter
- Install Claude Code — The terminal agent — and the moment your skill becomes operational. (~10 min)
- The Delegation Shift — From operating a tool to steering an outcome. (~5 min)
- Remote Control — Your phone as a window into the session running on your machine. (~10 min)
- Cloud Sessions and Dispatch — Work that survives your laptop closing — and the /teleport handoff. (~15 min)
If you take five things from this chapter
1. Pick the right home for the work. Remote Control when it needs your machine (files, local MCP, credentials); cloud sessions when it just needs to get done and survive your laptop closing.
2. Install knowledge, don't paste it. Skills in .claude/skills/ load themselves. If you're pasting the same context twice, you're doing it wrong.
3. Persist every correction on the spot. "Update the skill" / "add that to CLAUDE.md" — the distillation pipeline only works if corrections actually flow into it.
4. Measure direction-to-execution ratio. The Dispatch benchmark: 25 minutes of steering for 3+ hours of output. If you're watching Claude work, delegate bigger pieces.
5. /teleport is the handoff. Start work anywhere; pull the branch and full conversation into your terminal when it's time to finish properly.