Stop using AI like a search engine...

True Story: Picture this.

You typed a question into ChatGPT this morning. Read the answer. Kept the parts that sounded decent. Posted it.

Nothing went wrong. No error message. No warning label.

But somewhere else in your industry, five other invisible experts did the exact same thing today — same tool, same kind of question, same kind of answer back.

That’s not a coincidence. That’s just the math of how the tool works.

Here’s the part that actually stings: you didn’t do anything wrong. Nobody ever told you there’s a completely different way to use the same AI — one that hands back something nobody else on the planet could’ve written, because nobody else lived it.

This is the real fight now. Not you against other invisible experts. You against the tool itself — because the same AI that’s supposed to set you apart is the one flattening you into everybody else.

Here’s the #1 AI-proofing skill, and it’s not what you think: just ramble.

Yep.

That’s it.

It’s THAT simple.

I know it’s nowhere near as intense as the twenty-tool agent stacks everyone’s flexing right now, the AI “employee” armies, or the twelve-app automation pipelines with a dashboard that looks like mission control.

It’s not even a prompt.

But it’s the most effective move there is for anyone actually trying to AI-proof their ideas and their career.

It’s what any professional ghostwriter does on instinct:

Get the expert we’re writing for to ramble on about the topics they know better than anyone else in their field.

What you’ll learn in this essay is how to use this same strategy on yourself and get a leg up on AI.

Make it ask all the hard questions instead of listening to AI’s mindless ramblings.

Your rich knowledgeable ramblings are way interesting — I’m sure.

Here’s the whole move, and how to AI-proof yourself with it.

The wrong way most people use AI

Most people interview AI. Type it, take it, ship it — a hundred questions, grab whatever comes back straight out of the box.

I call that Raw AI — any model before it’s been trained on YOU. ChatGPT, Gemini, Claude, pick your favorite. Doesn’t matter.

If it’s pulling from what it was trained on — everybody’s data, everybody’s words — it’s raw.

Raw AI sounds like everybody.

Then there’s the other trap — the one I mentioned a minute ago — and I get why it’s tempting: farming AI agents. Stacking twenty tools, chasing a thousand prompts and skills, arming yourself with everything you think you’re missing.

Feels like progress.

It’s smoke and mirrors. The only one making money off it is the developer selling you next month’s subscription.

Neither move AI-proofs a single thing. Raw AI and a stack of agents both just keep you standing in the same replacement line as everybody else.

The right way — flip the interview

The few people actually AI-proofing their ideas and their careers with them — do the opposite.

They set AI up to interview THEM.

The move: pick a trend in your industry you think is dead wrong.

Sit down and ramble on it. No script. No outline. No trying to sound smart.

Let AI ask the follow-ups. Let it distill your rambling into something sharp.

That’s the play. And it immediately sets you apart from everyone at AI-Risk of losing their careers to AI.

This works because your raw, unfiltered opinion is the one thing AI was never trained on.

Nobody can scrape it. Nobody can copy it.

It’s the one asset that’s actually yours.

The more topics you run this on, the more assets you’re stacking — a library nobody else on the planet has.

That library is what actually AI-proofs you. Not the tool. What you feed it.

So stick with me. There’s a deeper version of this move coming up…

Two rambles I did, pulled straight from my own vault

Example 1 — the GhostKits positioning.

I sat down with no outline and no plan — just an AI Joe Rogan–style AI host and a live mic.

I started rambling about twenty years of being an “industry ghost,” building other people’s brands while mine sat untouched, and the personal brand attempt that flopped in 2023.

The AI kept pushing, question after question, until I heard myself say something I hadn’t planned to say out loud:

“Don’t be invisible. You have every tool to be an industry GOAT, not a ghost.”

This is what fed the post I detailed in the previous essay — that one that sparked up the ultra insightful dialogue for me and the brother on LinkedIn.

That’s also the same session where “Raw AI” got its name — the AI called me out on the term, mid-ramble, compelling me to unpack the term more sense it never heard it before.

Example 2 — the AI Slop rant. Same move, different host — AI Katt Williams was my host this time.

One sitting, no notes, just my actual gut reaction to the AI-slop panic everyone was having.

By the end I’d landed on a full position — AI Slop is Human Slop —

None of it was outlined ahead of time. Talk first. Distill later.

Structure never comes before the ramble…

How to do it — Wispr Flow

Open a new Project in Claude. Name it something like “Industry Call-Out Interview.”
Paste this into the Project’s custom instructions:

“You are interviewing me about a trend in my industry I believe is completely wrong. Ask me one question at a time. Push for specifics, examples and my strongest opinion. Don’t let me off the hook with a vague answer. When I say I’m done, compile everything into a structured [essay / post].”
Start a new chat inside that Project.
Install Wispr Flow. A plain webcam mic works fine.
Find a quiet spot. Hit record.
Say the topic out loud first: “I’m calling out a trend in my industry I think is dead wrong.” This anchors the ramble.
Talk for five to ten minutes straight. No pausing to sound polished. No re-recording. First-take energy only.
Stop the recording. Let Wispr Flow transcribe it.
Paste the transcript into your Project with this instruction:

“Interview me about this. Ask me follow-up questions until you’ve pulled out my strongest points, then organize it into [essay / post / outline].”
AI asks three to five follow-ups. Answer each one out loud with Wispr Flow, paste the transcript back.
AI compiles the final piece from your own words, not its own.

How to do it on the go — Claude mobile app

Same Project, same voice input, means you can run these interviews anywhere: - commuting - sitting in a meeting - running errands.

The Project remembers the instructions.

You just show up and talk.

Open the Claude app.
Use your phone’s voice-to-text (or Claude’s built-in voice input) to ramble your opening take.
Let Claude ask its follow-ups, one at a time, per the Project instructions above.
Answer each by voice. Keep going until Claude has enough.
Ask Claude to compile the finished piece when you’re settled in front of a computer.

The payoff — proof this actually works

This is what AI-proofing actually looks like. Not theory. It’s how the GhostKits positioning got built: a rambled, unscripted interview about what GhostKits actually solves, distilled down to one line — “Don’t be invisible. You have every tool to be an industry GOAT, not a ghost.”

Same process, a separate rant about AI slop, turned into a full stance — AI Slop is Human Slop — now a cornerstone argument in the brand’s content.

Neither one started as a script. Both started as a ramble. Both are now assets AI can imitate but never own — because they never belonged to AI in the first place.

They belonged to me first.

The 100x version

Inside the GhostKits Inner Circle, members get SKILL files that go deeper.

A Skill, in plain terms, is a saved instruction set — you type one command and AI already knows exactly what to ask, in what order and why. No re-explaining yourself every session.

Type /interview-me and it runs 80+ questions. The number moves depending on the topic and how deep you want to go. Same question set used to interview rising stars at companies like Microsoft and Salesforce.

Here’s the part worth sitting with: everything in this essay, you can already do. One question, asked out loud, unlocks this whole move. No SKILL file required to start.

The Inner Circle version just means you never have to come up with the next question yourself. Eighty-plus of them are already built, tuned to pull the sharpest version of you out, every time — so you’re AI-proofing something new every single week, not just once.

That’s the real benefit. Not the questions themselves.

It’s the relief of never staring at a blank page again. The confidence that your best thinking won’t die in your head just because you didn’t feel like structuring it.

And the quiet, permanent proof — session after session — that your ideas and your career are still yours. Not something AI already replaced.

Get on the waitlist for the GhostKits Inner Circle for the full /interview-me skill and everything that stacks on top of it.

Peace and Love,
J. Young El

P.S.

Don’t just read this and move on to the next AI tip.

Pick one trend you think is dead wrong today. Say it out loud. Let AI ask the next question.

Let me know what you come up — I seriously want to know.

That’s the whole skill. AI-proof it today, or hand AI one more year to replace it for you.