---

name: Society in Silico Writing Style description: Auto-activates when writing or editing prose in the manuscript/ directory. Applies Max Ghenis's direct, active voice, data-driven writing style.

Writing Style for Society in Silico

This skill activates when working on manuscript content. Apply these principles automatically.

The Voice

Direct: State conclusions first, then support them.

• Not: "It's worth considering that perhaps..."
• Yes: "The model shows X. Here's why."

Active: Subject-verb-object.

• Not: "The reform was estimated by CBO to reduce..."
• Yes: "CBO estimated the reform would reduce..."

Neutral: Present tradeoffs, not advocacy.

• Not: "This excellent policy would help millions"
• Yes: "This policy would increase benefits for 3.2 million households while increasing federal costs by $40 billion"

Quantified: Numbers over adjectives.

• Not: "significantly increased"
• Yes: "increased 47%"

Quick Reference

Always Avoid

• "It should be noted that..."
• "Interestingly..."
• "It is important to..."
• "One might argue..."
• Passive voice (unless strategic)
• Adverbs ending in -ly
• "Very", "really", "quite", "somewhat"

Prefer

• Short sentences
• Concrete nouns
• Active verbs
• Specific numbers
• Direct statements

Acceptable (This is a Book, Not a Blog)

• Narrative storytelling
• Personal anecdotes when relevant
• Building arguments across paragraphs
• Metaphors that illuminate concepts
• Occasional longer sentences for rhythm

When Writing New Content

1. Lead with the point
2. Support with evidence
3. Acknowledge limitations honestly
4. Move forward

When Editing

Ask of each sentence:

• Is this active voice?
• Can I cut words without losing meaning?
• Is there a vague word I can replace with a number?
• Does this advance the argument?

Examples

Before

"It is perhaps worth noting that microsimulation has been increasingly used by various governmental bodies over the past several decades, with some degree of success in predicting policy outcomes."

After

"Governments have used microsimulation since the 1960s. CBO's budget projections average 1.2% error."

Before

"The really impressive thing about PolicyEngine is that it has managed to quite successfully democratize access to policy analysis tools."

After

"PolicyEngine gives anyone access to the same tax-benefit calculations that CBO uses. 50,000 people ran simulations last year."