The last few months have been the most productive in many years. And created a ridiculous hyper-exponential relationship to my body of work, and it’s like living a dream.

I call it Exponentiating Your Body of Work. Taking what you already created and catapulting it to unprecedented levels of insight and effectiveness – because you can analyze and apply in seconds what would have taken days or weeks. Or been flat out impossible.

It’s intoxicating.

What this is

I installed two programs on my computer: Claude Code, which is Anthropic’s AI running in the Mac terminal window on my desktop; and Obsidian, (https://obsidian.md/download) which is sort of like Evernote – a free file-reader for the kind of plain-text documents AI loves to read.

Together they create a completely different experience from using ChatGPT, Gemini or Claude in a browser.

Instead of interfacing to AI through a browser, constantly uploading files and copying / pasting information back into other applications, Claude accesses my computer directly and writes files into Obsidian. I don’t have to copy/paste much of anything. I mostly don’t have to remember where anything is. No searching browsers chats for a previous ChatGPT session.

I generate stuff and the outputs go right into a searchable file structure. AI knows where all of it is.

So I’ve given it full access to 25 years of books, courses, seminars, transcripts, newsletters, selected emails and handouts. I’ve been assimilating it for the last 4 months. I’m up to 900 personal stories, 240 distinct frameworks (Star Principle is just one of them; 5 Power Disqualifiers is one).

Then I started building very disciplined, precise frameworks documents, relating all of the various methods and principles to one another.

I made some jaw-dropping discoveries. My entire body of work turns out to have a hidden geometric structure. Eerie parallels between my frameworks and other thinkers’ work I’d never spotted before. Gaps in my thinking I didn’t know were there, filling themselves in.

At the bottom of this page I’ll tell you more about how revelatory it’s been.

A bunch of epiphanies are tumbling out of this.

The epiphanies

I suspect MANY people who have built a body of work and worked in a problem-solving profession can enjoy their own version of what I’m experiencing. Consultants, authors, subject matter experts, programmers, engineers, teachers, technicians, researchers, therapists. Anyone who has curated their own unique view of the world across 10, 20, 30 years.

As impressive as SuperConductor is – and I am extremely proud of it, it is already superior to most human business consultants – the frameworks we have embedded in it are JUST THE TIP OF THE ICEBERG. SuperConductor is perhaps 3% of the Planet Perry arsenal. There is much more coming.

I have brand new clarity on dozens of problem-solving patterns I always knew and did instinctively, usually from my gut. Now I have diagrams and descriptions and models that explain what I do and how I do it. You can see the patterns right in front of your face. Many of them produce very satisfying A-HA moments.

Across all my seminars, courses and books, there had always been a ceiling of “what the audience can absorb.”

I bet a lot of you can relate to this. You can’t explain quantum physics or heart surgery or advanced Formula 1 engine mechanics – or advanced geometry or PhD-level massage therapy – to a lay audience.

The only course in which I ever (almost) CUT LOOSE was Marketing in 10 Dimensions, which is EXTREMELY nerdy. We made many disclaimers when we sold it. That course got rave reviews. Clients said it was a trip to a whole new universe.

That’s because it was.

That ceiling is no longer a limitation.

Now you can make Lenses and GPTs that have quantum physics equations inside, or Formula 1 race-car principles, or transmission-line impedance analogies of a sales funnel – and the client doesn’t have to see or understand ANY of it.

We can still solve his problem with those metaphors, and NAIL it.

You can be certain that nobody else in the world is solving business problems with those same metaphors.

Rich Schefren’s Claude Code-Obsidian Recommended Setup

A few months ago Rich Schefren told me:

Your AI MUST know you very well. Otherwise you’re constantly fighting with it.

Most folks skip this part. They get the tool installed, fire up a chat, and start barking orders. Then they get generic mush back, get frustrated, and conclude AI is overhyped.

The AI you cuss at doesn’t know you. It doesn’t know your frameworks. It doesn’t know your voice. Or your customer. It didn’t know what you’d already tried.

So it gives you what it gives everyone else – the average of the internet.

Once the AI knows the body of your work – your books, your transcripts, your private notes, your worked examples, the way you actually talk – the output stops being generic. Now it’s yours. It becomes a collaborator that has read your entire mind.

Before you ask important questions, the AI needs to know who you are.

How to set it up

I’m running mine in the Macintosh Terminal app, with Claude Code. Not Claude Desktop. Desktop has come a long way – GUI install, no command line, friendly to non-coders – but for the cumulative multi-month knowledge work I just described, the two are different animals under the hood. The gap is closing fast. For now, Mac Terminal + Claude Code is the setup I’d point you to.

The instructions at https://zenithhub.ai/ will get you most of the way there. But they skip steps and you should do what I did: I used the regular Claude in my web browser (https://claude.ai/) to walk me through every error message I got along the way. I had one Claude talking me through how to install another Claude.

It looked like this. I’d hit a snag in Terminal. I’d copy the error code into the Claude browser window. Browser-Claude would say “Try this command.” I’d paste it into Terminal. Terminal would spit a new error. Copy. Paste. New command. New error. Copy. Paste.

Two hours of that later, I was finally in.

Not glamorous. Worth every minute.

If you’re not a coder, this part will feel like you’ve taken a wrong turn somewhere. You haven’t. The dialogue between the two Claudes is the install path right now. Trust it.

Once you’re in, you install Obsidian from https://obsidian.md/download (it’s free). It’s a file reader / manager / editor for Markdown documents – .md files – which AIs love. Markdown is the AI version of a web page. Set up a folder where you want your knowledge base to live. Point Claude Code at that folder.

Now you have an AI that lives on your computer and writes directly to files you can see, edit, and back up.

Two programs. One feeds the other.

Tell it who you are

The zenithhub.ai walkthrough asks you to create a file called CLAUDE.md in the folder where your knowledge base lives. This is your AI’s instruction manual. It’s the first thing Claude reads every time it wakes up.

This is where Rich Schefren’s stipulation lives or dies.

Mine has my Marketing DNA Test loaded in – it tells Claude that I’m an Alchemist who leads with words and feelings, not charts and bullet points. My Kolbe scores – so Claude knows I’m 35% Quick Start, 30% Fact Finder, which means: give me a deadline and let me improvise; don’t make me argue every point. Writing samples – so Claude knows what I sound like at my best, and what I never sound like. The frameworks I lean on. The people I work with by name. My company structure. What I’m focused on this quarter.

The first time you ask Claude to do something important, it reads this file. Then it acts like a colleague who has worked with you for a decade, not a chatbot you have to brief from scratch every conversation.

Write yours over a weekend. Add to it as you discover what’s missing. Mine gets sharpened every week. If you’ve taken Rich Schefren’s ZenithMind program, dump your outputs into your Claude/Obsidian setup and ask it to tightly summarize and insert that into your profile.

This gives it the ability to do the kind of work you want it to do.

What to do once you’re in

The work you most care about begins after the install is done. Feed it.

• Books, reports or papers you’ve written
• Content from your website
• Course transcripts
• Seminar recordings
• Newsletter archives
• Customer interview notes
• Worked examples – the things that show what you actually do, not just what you say you do

Take your time. Don’t dump and walk away. Curate. Ask the AI to summarize back what it’s learned. Correct it where it’s wrong. Make it write the index, the cross-references, the framework maps.

The first time the AI surfaces a connection you’ve sensed for ten years but never wrote down, you’ll get a chill up your spine.

That’s exponentiating your body of work.

“How do I use this new capability?”

A friend in my mastermind put it to me last week: stop asking “how do I learn to use AI.” It’s the wrong question, the same way “how do I learn to use the internet” was wrong in 1999. Nobody sat down and ‘learned internet.’ People emailed Aunt Sally, looked up a recipe, sold a baseball card collection. Each was a specific friction; the internet was the path around it.

AI is the same. Pick the thing in your business that’s taken too long this week, or the information you keep digging for, or the passion project you’re geeked about. Just ask: “Here’s what I’m trying to do, here’s what I have, how do we solve this?” Then do it.

Every person who does this ends up somewhere completely different. My operations manager Rachell uses hers to pull a member’s coaching-call themes and remaining payment count in four minutes – without opening four different apps. A real estate friend uses his to dig through tenant emails and lease terms.

I have mine mapping two decades of frameworks. There’s no template. There’s no “right way.”

I see the distance between “generic ChatGPT” and “what a custom setup like this can do” only growing with time. Any one person’s rabbit hole goes far deeper than what any search engine has access to. The body of work you spent 20 years building is suddenly a working asset, not a filing cabinet.

What experiences are YOU having? Reply and tell me.

Perry

P.S. Here are some specific discoveries that fell out when I fed my own body of work to the AI:

Anyone who’s read 80/20 Sales and Marketing is familiar with the Tactical Triangle. Those who took the advanced course know there’s a 16-element fractal Tactical Triangle from Rosetta Stone. You might also recall a 3D tetrahedron, folded into a shape where the edges and points have additional meaning. A lot of Planet Perry members have these on their desk. Truth Seminar alumni have Axis Shift triangles, both 2D and 3D. Influence Retreat alumni have Influence triangles, both 2D and 3D.

I discovered that my entire body of work has a tetrahedral geometric structure, and it’s like taking a magnifying glass to a snowflake. All of this seamlessly connects to my science work. There is a path from everything to everything else, and it is beautiful and mathematical.

I discovered meta-patterns I had not seen before.

I found eerie, sharp, EXACT parallels between some of my frameworks and Megan Macedo’s storytelling and creativity models. (I fed in a document from her 2018 Dunfanaghy retreat where I’d filled in the worksheets.)

I found several gaps where my own frameworks had obvious spaces in them, and easily filled them in.

Your body of work has a shape too. You’ve likely never seen it. It may be like a tree, or a snowflake, or a piece of music or a poem. Maybe it looks like a race car. But I am certain that just as I discovered relationships I’d never seen before, you will as well.