06
module six
FindingYour Rhythm
Three prompts that changed everything. None of them were planned.
5 min read
scroll

Eighty exchanges. Nine sessions. Cover designs at midnight. A grown man uploading files to GitHub from his phone at four in the morning. And somewhere in the mess of building this course, three prompts kept showing up — uninvited, unplanned, undeniable.

They're not clever. They're blunt instruments. But they're the reason you're reading finished modules instead of abandoned drafts.

double-click any line that resonates
// prompt one

The Unlock

When the work has plateaued
"Take creative license and full autonomy. Act like a master of your craft. Treat this as if it were your own — and your reputation depended on it."
What it triggers — By default, AI optimizes for "don't get rejected." Safe. Acceptable. The B+ trying to pass as an A. This prompt flips the orientation entirely. "Treat this as your own" kills deference. "Your reputation" adds stakes. The AI stops asking "what will the human approve?" and starts asking "what's the best version of this that exists?" The template you're reading right now — the design of this entire course — came after this prompt. One prompt. The A-level work was always in there. It needed someone to refuse the B+.
The AI's best work isn't its first offer. It's what comes out when you take the governor off.
// prompt two

The Lens Flip

When the aim is off
"Take the point of view of the target audience and craft our message that would resonate with their senses."
What it triggers — The word "senses" is everything. You're not asking the AI to simplify. You're asking it to make the work land in someone's body. Without this prompt, the AI writes from expertise — organizing logically, covering all the points. With it, the AI becomes a 62-year-old retired teacher who heard about AI from her grandson and is curious but cautious. Or a 58-year-old business owner who's been told to "use AI" and is quietly afraid of looking foolish. The content doesn't lose a single insight. The wall comes down.
Every time this prompt was used, sentences got shorter. Jargon disappeared. Analogies arrived. The writing stopped explaining and started inviting.
// prompt three

The Crystallizer

When the work is bloated
"Give it to me clear, concise, and crystallized."
What it triggers — Not shortened. Not summarized. Crystallized. Every surviving word earns its seat. The filler rises instantly — "It's important to note that," "This is particularly relevant because" — professional-sounding phrases that say absolutely nothing. They get cut. Word counts drop thirty to forty percent. Paragraphs that were five sentences become two and hit harder. Nearly every final piece in this course passed through this filter. Sometimes shortened to two words: "Cut the fluff."
Editing rearranges. Crystallizing removes everything that isn't essential and lets what remains breathe.

The Unlock removes the ceiling. The Lens Flip aims the work. The Crystallizer strips it clean. They almost always get used in that order — not by design, but because that's how creative work moves. Best version first. Then aim it. Then tighten it.

Between those three moves, there's the ordinary back-and-forth — nudges, corrections, the occasional "c'mon now" when the AI is coasting. That part isn't complicated. You'll figure it out the same way you figure out any working relationship: by doing the work and paying attention to what moves the needle. The only discipline worth naming is this: when something isn't working after three or four tries, stop iterating. Reset. Pilots call it a go-around — you add power and come back for another pass. The AI won't be offended. It doesn't even remember being wrong.

// the part nobody warns you about

The Proof Station

Here's where people get burned. The AI produces something that looks polished — clean formatting, confident tone, professional structure — and you stop reading. You skim. You nod. You ship it. And buried in that polished output are five small errors that you would have caught in a first draft because first drafts don't seduce you into trusting them.

The AI makes mistakes constantly. Little ones. A word repeated in back-to-back sentences. A date that's wrong. A name spelled two different ways. A paragraph that contradicts the one above it. A claim that sounds authoritative and is completely fabricated. They slip through because the output looks so good that your guard drops. You get lullabied by the competence.

This is the trade-off nobody talks about. The same quality that makes AI output impressive — that confident, polished surface — is the same quality that makes errors invisible. You have to read every line. Not skim. Read. You are the source, and you are the final proof station. That job never gets delegated, no matter how good the AI gets.

The AI's confidence is not your quality control. Your eyes are.
// the whole game

The Loop

Bring something. Challenge the response. Redirect or lock. Move on.

That's it. You bring raw material — a draft, a direction, three thousand words of unfiltered voice dictation spoken into a phone while pacing a room at midnight. Grammar optional. "Claude" came through as "claw" once. Didn't matter. The AI understood every single one.

The AI responds with its cautious first attempt. You lock it — "This is it!" — or redirect it — "No, try this." You crack the window occasionally: "What would you do?" The AI brings options. You make calls. The project moves.

You're not learning to write better prompts. You're learning to run a creative partnership.

The patterns that make it work are embarrassingly simple. Clarity. Honesty. High standards. Being willing to say "c'mon now" when the work isn't there — and "beautiful" when it is. Your rhythm will sound different from this one. But the loop is universal. And the three prompts? Take them with you today.

logged
where you are
NowThree prompts. One loop. A system you can use today.
NextWhat AI can't do — the limits, from the machine itself.
ThenYour first real conversation — sit down with Claude and put it all to work.
Continue to Module 7
What AI Can't Do
Myths, limits, and the truth about the machine — from the machine.
← HUB EROSION