Crossing
the AI
Threshold.
The walkthrough that replaces a $100 live workshop.
Go outside.
Bare feet on the earth. 10 deep breaths.
I know this is weird for an AI tutorial. But you're about to spend a lot of time on your computer. This is the counterbalance.
3 minutes. I'm serious.
The tutorial will be here when you get back.
The most tangible line
I've ever crossed
in my life.
It feels like Bitcoin week one, but you already know it's going to $150K. Everyone still thinks AI is a chatbot you ask questions to.
What we're building today is a system that runs your entire operation. Persistent. Compounding. Yours.
You're about to walk through a doorway you'll never be able to walk back through.
300+ hours in.
Here's what
that looks like.
By the end of this,
you have this.
A working command center dashboard, open in your browser, connected to your files. Plus the foundation for a system that gets smarter every time you use it.
You're not using
a chatbot. You're
building a system.
Get an answer. Leave.
Three tabs.
Three different things.
Cowork isn't magic.
It's just a connected folder.
✓ Memory file found — reading context
► Ready. What would you like to build?
When you start a Cowork session, you pick a folder. That's your workspace. Claude can now read every file in it, create new files, edit existing ones, and remember what it did.
Every rule you save, every project you track, every piece of context — it lives in that folder as a real file. That's what makes it persistent.
Training wheels
vs the real bike.
Today we use Cowork because it's the fastest way to get you building. When it starts feeling slow, VS Code is the fix.
Friendly, fast, visual.
No VM, no friction.
How the system
actually works.
Four layers. Each one builds on the last. You don't have to manage these consciously — but knowing they exist changes how you think.
Four file types.
That's basically it.
You don't need to understand these to use them. Claude builds them. You just need to know they exist when you see them.
Anytime Claude does something you want it to always do, or you tell it a rule, say this. It writes it to a file. The file persists. Every future session reads it.
When I say "my podcast" I mean the Lucid Edge show.
My brand voice is warm, direct, never corporate.
Save all of that to your memory.
So. That's the most powerful phrase in this walkthrough. Use it constantly. That's how the system gets smarter every session.
You need all 5.
If you're missing any,
pause and handle it now.
Create your
CLAUDE folder.
Right-click your Desktop. New Folder. Name it CLAUDE. That's it.
This folder is your workspace. Everything Claude builds lives here. Every rule, every dashboard, every project. The folder IS the brain. Keep it on your Desktop so it's always easy to find.
Download the
master file.
Download the .md file from the link you received. Drag it into your CLAUDE folder. That's your blueprint. It tells Claude exactly what to build and how to set up your system.
When Claude reads this file, it knows your name, your goals, the structure you're building toward. It's not starting from zero — it's starting from you.
CLAUDE/
cat-master-prompt.md ← this file
(everything else gets created in the next steps)
Connect Cowork
to your folder.
This is the moment Claude gets access to your files.
Open Claude Desktop. Click the Cowork tab. Click “Select Folder” and choose your CLAUDE folder. It should show as connected right after.
Claude can now read what's in there and create new things. That green checkmark is the whole game starting.
Your first
prompt.
We'll be creating files and folders here. Let's stay organized.
Paste this exactly. Hit send. Claude will read the file and start asking you questions about your business, your projects, your goals.
Right after you send it, it'll probably say something like “I can see your master prompt. Let me ask you a few things.” That's correct behavior. Keep going.
Answer Claude's
questions.
Claude is going to ask about you. Your name, your business, your projects. Don't overthink this. Brain dump. Be messy. You can refine later.
The more context you give, the better your command center will be. Context is currency. You deposit now, you earn forever.
Build the
command center.
Build it as an HTML file I can open in my browser.
This might take 2–3 minutes. Claude is building your folder structure, your rules file, your naming conventions, and your dashboard all at once.
When it's done, it'll give you a file path. Something like /Desktop/CLAUDE/command-center.html. Find that file in your CLAUDE folder and double-click it.
Open it.
Welcome home.
Make your
first edit.
Go back to Cowork. Tell Claude to change something visual. Anything. This is the moment you feel the loop.
You talked. It recoded. The thing changed.
That's the entire game. Right? You just built something without writing a single line of code. Keep going.
Check your build.
By now you should have all three of these.
One hub.
Many views.
The command center is the map. Dashboards are the territory. Here's how the three layers relate.
So the command center has a widget that shows “3 active projects.” Click it and it opens your Projects Dashboard. That dashboard has cards. Click a card, it goes deeper. Claude builds all of it.
Your command center
looks good. Now make it
do things.
Your secret keyword: “function”
“I want my command center to have a function where...” and finish the sentence. Claude knows what function means. You don't need the technical word.
Talk to your computer.
Seriously.
You'll talk 5x faster than you type. Once you feel this, you won't go back.
Tell Claude what
you care about.
It follows forever.
Claude is your most literal employee. It needs explicit rules. But once you write a rule, it's there. Every session reads it. You said it once — you never say it again.
Session flow.
How every session runs.
Write this down: Handoff Report. When the chat gets slow, when Claude starts repeating itself, when responses feel thin — ask for a handoff. New session, paste it in, fresh context. That's the move.
You earn your
driver’s license
at 1,000 prompts.
That's when it starts to feel automatic. When you stop thinking about what to say and just talk. When you're two steps ahead because you know what Claude will need.
The fastest way to get there: ask every question that comes to mind. Use tokens now to learn. You'd rather learn early than fix things later.
An agent is just
a folder with rules in it.
When you hear “AI agent” or “AI employee” — all that means is: a folder with .md files that make Claude operate a specific way when you work in that folder.
Each agent has its own memory, personality, and scope. It doesn't know about your other projects unless you explicitly tell it to. That's by design.
MEMORY.md
rules.md
voice-guide.md
active-projects.md
Everything you build
should feed back
into itself.
Your rules get better as you use them. Your memory files grow. Your command center reflects reality. Your sessions get faster.
You refine once. It remembers forever. That's the compounding.
This is not a tool you open when you need it.
This is infrastructure your whole life runs on.
Your first week.
Day by day.
You're not the first
to cross this threshold.
Claude pricing.
What plan you actually need.
If you want to go
faster with help,
here's what's available.
DM me on Instagram or reply to any email. No pitch, no funnel. Just tell me what you're building and we'll figure out the right next step.
Keep
building.
Ask Claude first. Then ask my chatbot. Then ask me.
Every time something works: save that to your memory.
Go earn your first 1,000 prompts.
