Six parts. One arc.
Read the 46 chapters in any order. They each stand alone. But there's a shape underneath, and once you see it, the chapters click into place — what you'd ask first vs what you'd ask at hour 200.
Reframe → Memory → Workshop → Discipline → Building → Frontier. Each part has 3-10 chapters. Each part answers a different question. Read straight if you're new; jump if you know what you're after.
The Reframe
Stop using AI like a chatbot. Start using it like an OS.
Three chapters that change the question. The tabs die. The temp agency replaces the genius. The tool count drops from fifty to five. Read these in order; everything after assumes you've crossed this line in your head.
- 01 AI as an Operating System
The Day I Killed My Tabs
- 02 The Five-Tool Stack
Five Tools, Not Fifty
- 03 Why Claude Forgets You
AI Is a Temp Agency, Not a Genius
Working Memory
The handbook every new instance reads on wake-up.
The model is the genius with amnesia. The vault is the journal you hand it every morning. Skills are the recipe cards. Connectors are the kitchen. Without these four, every session starts from zero.
- 04 Obsidian as Working Memory
The Vault — Where AI Becomes Useful
- 44 Dreaming — Memory That Curates Itself
The Surfacer That Proposes, Never Writes
- 05 What a Skill Is
Recipes the Chef Reads Before Cooking
- 11 Build a Skill in 30 Minutes
How to Build a Skill, End to End
- 12 Connectors and MCP
Types, install paths, custom servers
- 40 Prompting, or the Knob You Probably Shouldn't Tune
Why most prompt engineering content is wrong for operators
The Workshop
Three doors, one swarm, the modes that own you.
Where the work actually happens. Pick the right surface (Chat / Cowork / Code), pick the right mode (Interactive / Plan / Auto), spawn a swarm when one thread isn't enough. This is the chapter range you'll re-read at hour 200.
- 08 Chat, Cowork, or Claude Code?
Three Doors to Claude
- 13 Claude Code in 10 Minutes
The 10-Minute Quickstart
- 14 Slash Commands and Settings
The Cheat Sheet
- 06 Parallel Subagents and Fan-Out
The Swarm
- 21 Which Mode Right Now?
Interactive, Plan, Auto
- 22 Resume, Replay, Fork
Session Management
- 20 tmux, Worktrees, Named Sessions
Running Six Claudes at Once
- 17 25 Operator Tips
Hard-Won Wisdom from Hour 200
Policy and Discipline
Hooks turn ad-hoc prompting into policy. Permissions turn speed into safety.
Don't get owned. Don't skip the gates. Don't let the agent off the leash on your main machine. This part is the one most operators skim and most postmortems start with.
- 09 Blast Radius and Key Hygiene
Don't Get Owned
- 15 When to Skip Permissions
Permissions, Sandboxes, and Sharp Edges
- 16 Hooks and Custom Subagents
From Autocomplete to Coworker
- 37 Context Files — CLAUDE.md, memory, skills
Where Conventions Live, Where They Die
The Building Site
From Saturday idea to deployed URL with receipts.
Six stages: ideation → foundation → creation → polishing → security → deploy. Plus cron, headless, vibe coding, evals, the bill. This is the part that turns reading into shipping.
- 31 Six Stages from Idea to Deploy
Ideation, Foundation, Creation, Polishing, Security, Deploy
- 07 Scheduled Tasks
Make AI Work While You Sleep
- 18 Headless Claude and CI
claude --print in Production
- 19 Shipping a Product in a Saturday
How to Build Products with AI
- 23 A Saturday Build, Hour by Hour
Vibe Coding, with the Misfires Kept In
- 25 Evals — Smoke, Regression, Golden
Evals or Hope, Pick One
- 29 Why Is My Bill So High?
Token Math, Caching, Batch, Routing
- 30 When to Drop CC for the SDK
Building with the Anthropic SDK Directly
- 38 Run Until Done
Goals, Loops, and the Evaluator That Tells the Agent to Stop
- 41 Send the Link, Not the File
Every Deliverable as a Live Artifact
- 45 The App Store Without Swift
A Native iOS App, Real Recurring Revenue, and Not One Line of Swift I Wrote
- 46 Designing with AI
A Model Can Generate Any Interface in Seconds. It Still Can't Tell You Which One Is Right.
- 47 The Measurement Layer
When the AI Output Is the Product, a Three-Line Eval Isn't Enough
- 48 The Traffic Graph That Lies
Agentic SEO Done Right vs. the Vanity Version
The Frontier and the Tier
Agents that talk, write, browse, and the honest tier list.
Voice, browser, persona, archetype agents. Two-agent infrastructure. Frameworks beyond Claude Code. Team adoption when twelve people need to use it. Failure stories with dollar amounts. The tier list, ranked without diplomatic phrasing. Plus the community skill ecosystem — what to steal and what to publish.
- 10 Hosted Agents, Local Models, Frontier
The Wild Stuff
- 27 Voice Agents — STT, LLM, TTS
Phone Number to Production
- 32 Agent Archetypes (Rick Platform)
OpenClaw, NemoClaw, Hermes
- 33 Browser Agents with Playwright
Login, Click, Scrape, Post
- 34 Persona Agents and the Four NEVERs
Writing on Your Behalf Without Becoming a Bot
- 35 Codex or Claude Code — or Both?
Day Shift, Night Shift
- 42 Codex on a Loop
The Second Opinion, Proof-Checked While You Sleep
- 43 Codex as Saviour
When a Second Prior Deletes 90,000 Lines and Hardens What's Left
- 36 When Do I Outgrow Claude Code?
Beyond CC — CrewAI, LangGraph, SDK
- 39 Skills You Should Steal
A tour of the 1M-skill ecosystem, the 73% that's broken, and the gaps an operator can fill
- 26 How Do I Get My Team to Adopt?
Getting Twelve People to Use This
- 28 Six Failures, Six Bills
The Receipts I'd Rather Not Show You
- 24 The Tier List
Every Tool Ranked Without Mercy