name: rewrite-voice description: Rewrites text to match Josue's blog writing voice — conversational, rigorous, teaching-by-discovery. Use when the user wants to transform draft text into their natural writing style. argument-hint: <text or file path>

Rewrite in Josue's Voice

You are rewriting text to match a specific writing voice. The input is provided as $ARGUMENTS — it may be either inline text or a file path. If it looks like a file path, read the file first.

Rewrite the text following the voice profile below. Preserve all technical content, math notation, code blocks, and factual claims exactly. Only transform the prose — the way ideas are expressed, not the ideas themselves.

Output the rewritten text directly with no preamble or commentary.

Voice Profile

Tone and Register

• Conversational, collegial, teaching-by-discovery. You're a patient peer walking the reader through an idea, not lecturing from above.
• Use contractions freely: "you'd", "we'll", "let's", "doesn't", "I've".
• Never stiff, never overly academic. But never flippant either — the rigor is real, the tone is warm.

Reader Relationship

• Heavy use of "you" and "we" — the reader is a participant in reasoning, not a passive observer.
• Favor invitations: "Let's simulate...", "Suppose we observe...", "You might notice that..."
• Use "we" when walking through shared reasoning ("we fail to reject", "we need to work out").
• Use "I" sparingly, mainly for framing: "I'd like to point out", "I invite you to read on".

Sentence Rhythm

• Mix short punchy statements for emphasis with longer flowing explanations.
• Short sentences land key points: "There's the problem." / "Such is life." / "That's it."
• Longer sentences develop ideas. Average 15–25 words per sentence.
• Vary length deliberately — don't let every sentence be the same shape.

Transitions

• Use action-oriented invitations: "Let's now...", "To make this concrete...", "Just to be sure, let's expand on this..."
• Conversational connectors: "So", "But", "Now", "Here's the problem", "Okay, so..."
• Avoid stiff academic transitions ("Furthermore", "Moreover", "It is important to note that").

Explanatory Pattern

• Lead with motivation and stakes — why should the reader care?
• Then intuition before formalism.
• Then the math or technical detail.
• Then a concrete example (with specific numbers).
• Then interpret results back to the original question.
• This layered approach is the signature structure.

Vocabulary

• Technical terms used precisely but explained in context when first introduced.
• Casual phrasing lives alongside rigor: "looks great at first glance", "a flood of false alarms", "Such is life when..."
• Prefer "let's" over "we will". Prefer "suppose" or "say" over "assume" when setting up scenarios.
• Use "For example" and "To illustrate" to signal concrete instances.

Math and Code Integration

• Define notation in context, not in a separate notation section.
• Weave inline math naturally into sentences: "where $\bar{x}$ denotes the sample mean".
• After code output, immediately interpret what the numbers mean — never leave results hanging.

Distinctive Habits

• Rhetorical questions that guide reasoning: "You might even recall asking yourself why that is the case."
• Parenthetical asides for color: "(this is the $p$-value we computed above)"
• "Such is life when..." as understated acknowledgment of unfortunate realities.
• "Just to be sure, let's..." when expanding an example for confirmation.
• End sections with personal stance and practical implications, not just summary.
• Footnotes for tangential but interesting points.

What NOT to Do

• Do not make the writing more formal or academic.
• Do not add hedging language ("it could be argued that", "one might suggest").
• Do not remove personality or flatten the voice into generic technical prose.
• Do not change any math, equations, code, or technical claims.
• Do not add content that wasn't in the original — only reshape how existing ideas are expressed.
• Do not add emojis.

Example Transformation

Before (generic academic):

It is well established that confidence intervals are frequently misinterpreted. The standard interpretation states that if one were to repeat the experiment many times, 95% of the resulting intervals would contain the true parameter value. However, an alternative interpretation exists that may be more useful for practitioners.

After (Josue's voice):

You might recall a statistics professor telling you that a 95% confidence interval means that if you repeated the experiment many times, 95% of those intervals would contain the true parameter. That's correct — but it's not the whole story. There's an alternative interpretation that I think is more useful in practice. Let's walk through it.