Daniel J. Dick

Disclosure: This page contains affiliate links. If you purchase through them, we may earn a commission at no extra cost to you.

How This Article Came To Be

Recently, I did something I often encourage other people to do but don’t always do for myself:

I asked for help.

I took a sample article I had written on free speech and censorship and asked an AI assistant,

plugged into my script-based ghostwriter system, to:

• Analyze my writing style.
• Point out my strengths and weaknesses.
• Suggest the 20% changes that would produce 80% of the improvement.
• Show how we could bake those improvements into my client configurations so future articles

come out stronger by default.

The feedback was specific, practical, and frankly, humbling—in a good way. It also gave me a

clear path to help future clients get the same kind of boost.

This article summarizes the core advice the AI gave me and how I’m now using it inside my

`clients/` configurations to generate better content for myself and, eventually, for people

I consult with.

What the AI Saw in My Writing

The AI started by acknowledging that some important things were already working in my sample piece:

• Clear emotional stance and urgency. The reader knows what side I’m on and why it matters.
• Strong use of repetition and rhythm. Sections like “We fight…” build momentum.
• Moral framing. Freedom is treated as an ethical duty, not just a preference.
• Direct engagement. Questions like “What do you think?” and “How will you fight?” invite action.

In other words, the voice is there. The problem isn’t lack of conviction; it’s structure and clarity.

The AI then focused on a small set of changes that would deliver outsized gains.

1. Replace Vague Accusations With Concrete Examples

One of the biggest weaknesses in my original article was vagueness. I wrote lines like:

> Big Tech, the FBI, Mainstream Media, and Social Media Take Action!

>

> Twitter, Facebook, and other social media began censoring.

Those sentences gesture at real concerns, but they don’t tell the reader what exactly happened.

There are no dates, no policy names, no specific incidents to look up or verify.

The AI recommended that for every strong claim like “they censored” or “they jailed people,” I

should force myself to answer a few questions:

• When did this happen?
• Where did it happen?
• Who was involved (platforms, agencies, people)?
• Under what policy, rule, or event did it happen?

That one habit—adding named, dated, specific examples—instantly makes a piece feel more like

real journalism and less like a rant.

2. Shorten Long, Tangled Sentences

My opening sentence used to be a marathon:

> What do you do when your liberators become your captors when your freedom to speak is replaced by censorship and other people… decide what ideas you can and cannot say, think, write, read, hear, support, or believe?

The idea is strong, but the delivery is hard to follow. The AI recommended a simple rule:

• Aim for 15–20 words per sentence on average.
• Break complex thoughts into 2–3 short sentences instead of one long one.

For example, the AI showed how I could rewrite that opening into a short series:

> What do you do when your liberators become your captors?

>

> When “freedom of speech” turns into quiet, private censorship?

>

> When people you never voted for decide what you can say, read, or even ask?

The content didn’t change. The structure did. That alone makes the writing easier to read,

more quotable, and more shareable.

3. Make Repetition Advance the Argument, Not Just Echo It

Repetition is one of my natural tools. I like lines like:

• “Freedom is something we must exercise.”
• “Freedom is something we fight for.”

The AI didn’t tell me to stop repeating; it told me to upgrade my repetition.

When you repeat a phrase, each repetition should add a new dimension:

• Ethical: Why is this right or wrong?
• Practical: What does this mean day to day?
• Historical: Where have we seen this before?
• Personal: What does this cost or require from us?

So instead of 4 near-identical bullets starting with “Freedom is something we must exercise…”,

I can keep the rhythm but make each line do new work. That’s a very small change in process

with a big payoff in impact.

4. Add Clear, Practical Action Steps

My original article ended with a strong question:

> How will you fight for freedom today?

But I left readers hanging. The AI pointed out that most people need concrete next steps.

So it suggested always ending with a short “What you can do now” section:

• 3–5 practical, realistic actions.
• Specific enough that someone could literally do them this week.
• Calibrated for the audience—no fantasies, just real moves.

For example, in the free speech context, that might mean:

• Read the actual “community standards” of one platform you use.
• Support one organization that defends viewpoint diversity.
• Have one honest conversation with a friend who sees things differently.

It sounds simple, but consistently adding this section turns a piece from pure commentary

into a call to informed action.

5. Steelman Opposing Views Before Critiquing Them

Another big upgrade was how I handle disagreement. My early drafts tended to describe opponents as

“leftists” or members of a “political correctness crowd” without really engaging with their best

arguments.

The AI recommended a discipline taken from good debate and good journalism:

• Spend one short paragraph stating the best opposing argument as fairly as you can.
• Only then explain why you disagree, using evidence and reasoning.

This doesn’t dilute conviction. It increases credibility. Readers can see that you understand

what the other side is saying and still find it lacking.

Baking These Improvements Into the Ghostwriter System

The most powerful part of this exercise is that it doesn’t end with one human edit. Because I’m

using a script-based ghostwriter system with per-client configuration in `clients//`, we can

encode these habits directly into the config.

For example, for one of my sites we:

• Updated the brand voice in `intake.yaml` to:

– Favor 15–20-word sentences.

– Use 1–3 sentence paragraphs with frequent headings and bullet lists.

– Require concrete examples for strong claims.

– Ask for a short, fair statement of the opposing view.

– Require a “What you can do now” section with 3–5 action steps.

• Tightened the content goals so target word counts are realistic and focused.
• Added explicit writing_habits constraints to remind the model (and me) to:

– Limit rhetorical questions and follow them with clear answers.

– Avoid overusing the same sentence stem.

– Prefer named events, dates, and policies over vague labels.

• Updated the prompt templates in `prompts.yaml` so every generated article:

– Starts with a concrete story.

– Includes at least two named, checkable examples.

– Contains a structured counterargument section.

– Ends with practical actions.

The result is a system that doesn’t just generate content in the abstract. It generates content

that sounds like me on my best days—only more consistent.

How This Can Help Future Clients

If you’re a future client reading this on my site, here’s what this means for you.

When I set up a ghostwriter configuration for your site:

1. We start with your real writing.

– You give me sample articles, emails, or reports.

– The system and I analyze your voice: strengths, weaknesses, and patterns.

2. We design a style guide that plays to your strengths.

– If you’re great at storytelling but weak on structure, we build prompts that

enforce better structure.

– If you’re strong on data but dry, we add narrative and moral framing.

3. We encode those habits into your client’s `intake.yaml` and `prompts.yaml`.

– Tone, target audience, constraints, and content goals all become explicit.

– The AI stops guessing and starts writing according to your house rules.

4. We use the script-based menu system to generate and refine content.

– Seed topics.

– Generate drafts.

– Run policy checks.

– Upload drafts to WordPress as safe, reviewable content.

5. Over time, we can add a “review and rewrite” pass.

– The system can scan older posts.

– Identify which ones most need a rewrite.

– Generate rewrite drafts (`-rw`, `-rw2`, etc.) that apply your latest best practices.

In short: this isn’t just generic AI writing. It’s a customized companion that learns

how you want to think, write, and persuade—and then helps you do that more consistently at scale.

Closing Thoughts

When I first asked the AI for feedback on my writing, I expected a few generic tips.

Instead, I got a compact set of habits that, if I practice them, will make almost

everything I write clearer, more concrete, and more useful to readers.

By wiring those same habits into my ghostwriter configurations, I can:

• Improve my own work.
• Offer a more powerful writing and content strategy service to clients.
• Eventually roll these improvements into the web-based system so more people can

benefit from the same process.

If you’d like help doing something similar for your own site—analyzing your writing,

upgrading your habits, and encoding those upgrades into a repeatable system—that’s

exactly what this AI-powered ghostwriter setup is designed to do.

---