Jonathan Haaswritingnowusesabout
emailgithubx
Jonathan Haaswritingnowusesabout

When AI Learns to Write Like You: A Meta-Analysis

June 19, 2025·2 min read

I asked Claude to analyze my writing style across my blog posts. The patterns it found -- and the ones I didn't know I had -- were genuinely surprising.

#ai#writing#developer-experience#meta

I asked Claude to analyze my writing style. No specific instructions. No examples of what to look for. Just "deeply examine my blog posts to determine how I write."

It produced a 363-line style guide that documented patterns I had never consciously noticed.

The Unconscious Patterns

Structural: I open with personal anecdotes or bold claims almost every time. Paragraphs rarely exceed four sentences. Rhetorical questions serve as section transitions. Strategic fragments for emphasis.

Voice: Direct address ("you") throughout. Present tense. Contractions everywhere. Zero hedge words -- no "maybe" or "perhaps."

Formula: Hook with controversy or experience. Identify the problem. Promise value upfront. Build tension between conventional wisdom and reality.

Claude also identified what I never do: no academic paragraphs, no passive voice unless forced, no over-explaining, no apologies.

This post follows those same patterns. That is the meta layer -- using extracted guidelines to write about extracting guidelines.

Why This Is Useful

Consistency at scale. When you create content regularly, maintaining a consistent voice is hard. Documented patterns let you check whether a draft "sounds like you" against measurable criteria instead of gut feel.

Delegation without dilution. Whether you work with AI or human writers, shareable guidelines maintain your voice. My extracted patterns are now in my CLAUDE.md file. Every AI interaction starts with context about how I write.

Deliberate pattern-breaking. Understanding your defaults lets you choose when to break them. I discovered I default to physics metaphors for abstract concepts -- technical debt as entropy, architecture as gravitational forces. Now I can lean into that tendency or deliberately reach for something different.

What It Cannot Tell You

The analysis captures how I write but not why. The patterns emerged from years of iterating on what resonated with readers. The AI identifies the patterns but not the reasoning behind specific choices.

There is also an authenticity question. When AI can replicate your style, what makes something "yours"? The answer is intentionality. The AI can mimic patterns. The decision to use those patterns -- or break them -- remains human.

Style extraction is a tool for self-awareness, not a replacement for voice. The experiment takes 15 minutes to set up against your own writing. You will likely discover habits worth keeping and habits worth breaking.

share

Continue reading

The CLI Renaissance: How AI is Driving the Command Line Revolution

AI coding assistants output shell commands, not GUI instructions. That single fact is reversing a decade of developer tooling trends.

Building for Humans AND Machines: The Dual-Audience Problem

Every web design decision now must serve two audiences: humans who browse visually and AI agents that consume data programmatically. The architectural...

AI Code Review Is Reasoning, Not Pattern Matching

AI code reviewers moved from rules-based checking to reasoning-based analysis. The gap between what they catch and what humans catch is closing fast.

emailgithubx