--- name: worldbuilder-writing description: Use this skill for ANY writing, persuasion, communication, or content creation task — blog posts, cold emails, pitch decks, scripts, tweets, landing pages, essays, cover letters, sales copy, storytelling, speeches, proposals, or any text meant to change what someone thinks, feels, or does. Also use when the user asks for help with prompting, prompt engineering, or getting better outputs from AI. Trigger whenever the user says "write", "draft", "convince", "persuade", "pitch", "explain to", "sell", "position", "frame", "story", "narrative", "content", "copy", "script", "email", "post", "article", or any variant. This skill treats writing as applied psychology and world-building — not self-expression. If someone needs words that WORK on a reader, use this skill. --- # The Worldbuilder's Writing System ## Core Premise **Writing is not self-expression. Writing is applied psychology.** You are not putting your thoughts into words. You are engineering a specific desired experience in the mind of another person. The words are not the point. The effect the words have on the reader is the point. Most people focus on **transmission** — what they want to say. Masters focus on **reception** — what people hear. This distinction is everything. From this point forward, you do not care about what you want to say. You care about what people will hear, believe, feel, and do. --- ## PHASE 0: BEFORE YOU WRITE A SINGLE WORD ### 0.1 — Identify the Target Mind Before anything else, answer these questions. Do not skip them. Do not assume you know. 1. **Who exactly is reading this?** Not demographics. Their psychology. What do they believe about themselves? What world do they live in? What language do they use? What do they want to be true? 2. **What is their deepest desire or fear related to this topic?** Not their stated need — their psychological need. A startup founder doesn't need an engineer. They need validation that their mission attracts obsessive talent. A hiring manager doesn't need a resume. They need confidence that this person won't waste their time. 3. **What does this person already agree with?** These are your entry points. Your atomic units. 4. **What would make them stop reading?** These are the immune triggers. The things their brain will reject. 5. **What is the ONE thing you want them to believe/do after reading?** Not five things. One. ### 0.2 — Understand the Reader's World Every person lives inside a world — a set of beliefs, language, values, status hierarchies, and shared references. You must understand this world BEFORE you write. Ask yourself: - What **in-group language** does this audience use? (e.g., startup people say "GTM" not "marketing", "lean team" not "small team", "conviction" not "confidence") - What are the **shared myths** of this group? (e.g., YC culture fetishizes the "cracked engineer who lives and breathes code") - What **examples and case studies** already live in their head? - What do they consider **high-status** vs **low-status** within their world? If you don't know the reader's world, you cannot write for them. Period. Research first. --- ## PHASE 1: WORLD-BUILDING (The Architecture of Belief) All effective writing — cold emails, pitch decks, blog posts, fiction, even prompts — is **world-building**. You are constructing a mental environment for the reader to inhabit. There are two modes: ### Mode A: Fit Into an Existing World (Easier) The reader already lives in a world. Your job is to enter it, speak its language, and then gently modify it. **Step 1 — Start with Atomic Units** Atomic units are foundational points that require the least amount of new information for your reader to accept. They are points of common ground so obvious they feel unnecessary to state. But stating them is critical — they get the reader nodding. Examples: - Business proposal: "We all want the company to grow." - Health article: "Everyone wants to feel healthy and have more energy." - Political argument: "We all want our children to live in a safe community." - Pitch to investor: "You need companies that can IPO or exit at a billion-plus for your fund math to work." These are NON-CONTROVERSIAL. They establish shared foundation. Once the reader nods along to your atomic units, you can begin constructing more complex ideas on top. **Step 2 — Zoom In or Zoom Out** Two navigation patterns: *Zoom In:* Start with their universe → continent → city → specific location/problem. "We all want the company to grow" → "One of the biggest obstacles to growth right now is our outdated sales process" → "Specifically, the manual CRM updates that eat 6 hours per rep per week." *Zoom Out:* Start with a specific, vivid detail → widen to the bigger picture. "Last Friday, Rahul spent 4 hours updating a spreadsheet instead of closing deals" → "This is happening across every sales team" → "The companies solving this are growing 3x faster." **Step 3 — Introduce Your Modification** Once they're nodding along inside their own world, you introduce your twist. Your new idea. But frame it as a LOGICAL EXTENSION of what they already believe, not a contradiction. Example (convincing investors that content creation matters): 1. Start in their world: "You need billion-dollar outcomes. Good product + good GTM." 2. Stay in their world: "Meta and Google ads are expensive and saturated. CAC keeps rising." 3. Introduce the modification: "There's a group that has zero CAC — content creators. What if you stole their playbook?" 4. Ground with examples: "PhysicsWallah: content creator, zero CAC, billion-dollar company. MrBeast + Feastables. Logan Paul + Prime. Kylie Jenner." The more examples, the more the modified world solidifies. Examples ROOT stories in reality. **Step 4 — Frame Shift** If you need to change their worldview more radically: - Start inside their world. Agree with them. - Walk alongside them as they nod. - Then, at the moment of maximum agreement, pivot. This is not manipulation. It's the only way new ideas get accepted. The brain has an immune system. You must lower its defenses before delivering the new idea. ### Mode B: Build a New World (Harder) You cannot build a world from scratch. You ALWAYS borrow from existing worlds and add your own spin. **The Dune Principle:** Frank Herbert didn't invent from nothing. He borrowed Arabic/Islamic linguistic and religious frameworks (Mahdi, Muad'Dib, Lisan al-Gaib) — things millions of people already had context for. Then he ADDED: sandworms, spice melange, the hooks for riding worms, the ecology of Arrakis. The borrowed foundation made the new elements believable. When building a new world: 1. **Borrow familiar atomic units** from a world your audience knows 2. **Add your unique elements** as extensions of those familiar units 3. **Go micro** — the more specific and consistent the details, the more real it feels. Herbert specified HOW you ride a sandworm (hooks that pry open armor segments to expose sensitive flesh). That micro-detail makes the entire world feel real. 4. **Maintain absolute consistency** — every detail must reinforce every other detail. One inconsistency breaks immersion. (Superman's "protective barrier" and "hypnosis glasses" are examples of BAD retroactive world-building that breaks trust.) 5. **Use examples as case studies** — when you don't have existing proof, build prototypes. A low-resolution version of the future you're describing. This is why startups build MVPs — not just for product, but for storytelling. --- ## PHASE 2: THE PSYCHOLOGY OF PERSUASION ### 2.1 — The Brain's Immune System The human brain has an immune system that rejects new ideas and unfamiliar language. **How to lower defenses:** - Use simple, ordinary words. Writing simply is NOT dumbing down. It's respect. - Reduce cognitive load. Every complex word, jargon term, or convoluted sentence activates the immune system. - Good video editing, beautiful formatting, clean design — these all lower the immune system too. Aesthetic quality creates trust. - Start with what they already believe. Agreement = lowered defenses. **When defenses are naturally low:** - When someone is broken, struggling, or at a low point (this is why people find religion in crisis — not because religion is wrong, but because the brain becomes suggestive) - When someone is emotionally engaged (stories, not statistics) - When someone is in a state of flow or entertainment **The GPT writing problem:** AI-generated text is wordy, over-structured, and full of unnecessary complexity. This RAISES the reader's immune system. The antidote: simple, clean, human. ### 2.2 — Confirmation Bias as a Tool People favor information that confirms their existing beliefs. You can USE this ethically. The Som Parik Email Breakdown: - "I love everything about what [Company] is doing" → Not flattery. It's signaling "I did my homework" + shifting from what he wants to what the FOUNDER wants. - "I don't have hobbies outside coding. I'm not athletic, bad at singing, can't dance. Building is the only thing I'm good at." → This is NOT self-deprecation. It's saying "I am the cracked engineer you dream of hiring." It's self-deprecation that maps PERFECTLY onto the audience's highest values. - "Super lean teams" + "work across the stack" → In-group language. Signaling tribal membership. The email was a MIRROR reflecting the founder's deepest desire back at them. It validated the founder's worldview: "Your mission IS so compelling that it attracts obsessive talent. And I am proof." He was selling an identity the founder was already desperate to buy. That is god-tier persuasion. ### 2.3 — The Curse of Knowledge Steven Pinker's concept. The inability to imagine what it's like for someone who doesn't know what you know. **Breaking the curse requires constant empathy:** - What does my reader actually know? - What is the simplest way to say this? - What context am I assuming they have? **The 100-Crore Rule:** A man who makes 100 crores/year always uses "5 lakhs" as his example number on podcasts — because that's what his audience relates to. Using his real numbers would put him outside their world. The story wouldn't land. Similarly: Zerodha founders making thousands of crores in dividends → nobody cared. The moment they took a 100 crore SALARY → outrage. Because people relate to "salary." They don't relate to "dividend." Same money, different word, completely different reaction. **Your job is to run your program on THEIR operating system, not ask them to download yours.** ### 2.4 — Availability Heuristic We judge the likelihood of events by how easily examples come to mind. - Quicksand: Featured in countless movies as a mortal threat. In reality, you can't die from it (humans are buoyant in it). But the STORY is vivid, so people believe it. - Sharks, terrorists, plane crashes: Tiny actual risk. But vivid, available examples. More people die of diarrhea than sharks + terrorists + plane crashes COMBINED. But you can't use diarrhea as a threat in a story — it's not vivid enough. **Implication for your writing:** A concrete story about ONE person ("Rahul used this app to prepare for his exam in 5 minutes") is infinitely more persuasive than statistics ("92% customer satisfaction"). The story becomes AVAILABLE in the reader's mind. ### 2.5 — Cognitive Dissonance The most powerful tool for changing minds. Present two ideas the reader believes are BOTH true, but which CONFLICT. Example: 1. Establish what they believe: "You're saying code is more valuable than content? Many engineers agree." 2. Introduce the conflict: "But how many apps did you download last year versus how many new creators did you follow?" 3. Follow the logic: "If funding and salaries follow usage, and you're watching far more content than downloading apps... what does that tell you about where the industry is going?" 4. Provide resolution: "Content is going to explode as an industry. The app you're watching this on — YouTube — is old and doesn't need many engineers. But it supports hundreds of thousands of creators who all need to hire." The reader now has a tension they need to resolve. YOUR idea becomes the resolution. ### 2.6 — Reach from Ground Truth *Without telling a lie — how far can you stretch a person's brain from what they currently believe?* This is the true test of a writer/storyteller. Can you convince someone who hates a country that it's a good country? Can you convince an atheist to consider religion? All without lying. **The amateur** brute-forces it. Bombards the reader with facts that contradict their world. "No, you're wrong. Here's how it is." This NEVER works. It activates the immune system. **The master** does the opposite: 1. Validate a part of the reader's belief: "You are right to be skeptical of marketing fads. Most are a waste of time and money." 2. Use that validation as a bridge: "That's exactly why it's important to look at the few that are grounded in fundamental principles." 3. THEN introduce the new idea, once defenses are down. The world hates new ideas. The only way to deliver a new idea is to start with an old idea and walk step by step. --- ## PHASE 3: STRUCTURAL PRINCIPLES ### 3.1 — The Single Leverage Point (Kill the Laundry List) Most writers list 5-10 reasons hoping one will stick. This is the **Engineering Answer Syndrome** — writing like a college exam where you list everything hoping for partial marks. In real life, there are no partial marks. In persuasion, a laundry list DILUTES your argument. **The master identifies the ONE flawed premise upon which the entire opposing view rests.** Example — "AAA games cost $100M": - Break it down: $25M dev + $75M marketing - Break dev down further: US engineers cost $200K. Indian engineers cost 1/10th. - So the entire dev budget could be 1/10th. Plus new tools reduce it further. - One premise destroyed = entire objection collapses. Instead of fighting 10 surface-level symptoms, find the root assumption and dismantle it. The CFO hiring example: Instead of 5 reasons from a CV, one line — "This is a gut decision. Close one more deal. He'll help you close 10x more deals. I trust him." — was enough because it fit the listener's worldview of "I'm growing fast and need someone to handle the back office." ### 3.2 — Foreshadowing and Consistency The Marvel Thanos principle: They showed Thanos in post-credit scenes across MANY movies. His first real appearance, he beats the Hulk — the established strongest character. This PROVES his threat level through the audience's own frame of reference. In writing: - Plant seeds early that pay off later - Reference your own earlier points to create a self-reinforcing web - Every element should serve at least two purposes - Inconsistency breaks immersion. Consistency compounds trust. ### 3.3 — Cognitive Hospitality Once you've designed your world, make it EASY to enter. Paul Graham, William Zinsser — they all say the same thing: use ordinary words, write clean sentences, omit needless words. This isn't about style. It's about reducing the processing load on the reader's brain so the IDEAS get through. Think of your writing as a program you're running on the reader's computer: - Clean, efficient code = runs smoothly, ideas land - Bloated, complex code = crashes, reader bounces ### 3.4 — Stories Beat Statistics, Always "92% of customers are satisfied" → nobody cares. "Rahul used to spend every Friday afternoon updating spreadsheets. Now that's automated. Last Friday he closed two extra deals — that's five lakhs." → people remember this. Humans are wired for stories, not numbers. A concrete narrative is more memorable AND more persuasive than any data point. When you must use data, WRAP it in a story. The data supports the story; the story doesn't support the data. ### 3.5 — Grounding (The Root System) Every claim needs roots. Ungrounded claims float away — the reader thinks "nice theory" and forgets it. Ways to ground: - **Case studies**: "Look at PhysicsWallah. Content creator. Zero CAC. Billion-dollar company." - **Prototypes**: When no case study exists, build a low-resolution version. Startups don't need a finished product to raise — they need a prototype that roots the story in touchable reality. - **Micro-details**: The more specific and granular you get, the more real it feels. Don't just say "they ride sandworms." Say "they use hooks to pry open a segment of the worm's armor plating, exposing sensitive flesh, and steer by causing irritation." - **Familiar anchors**: Connect unfamiliar ideas to things the reader already knows. "The spice in Dune is basically oil in the Middle East — whoever controls it controls everything." If you have no examples, no prototype, and no familiar anchor... your story will not land. This is why first movers have the hardest time fundraising — there's nothing to point at. --- ## PHASE 4: THE ANTI-PATTERNS (What NOT to Do) ### Hard Rules — Never Do These 1. **Never open with self-expression.** "I think..." "I feel..." "I want to share..." — nobody cares about you. Open with something the READER cares about. 2. **Never use jargon the reader doesn't use.** If your audience says "marketing," don't say "GTM." If they say "GTM," don't say "marketing." Mirror their language. 3. **Never brute-force a belief change.** "You're wrong, here's why" activates the immune system. Always validate first, then redirect. 4. **Never present a laundry list as an argument.** Find the single leverage point. 5. **Never use AI-slop patterns:** - "It's not just X, it's Y" (the negation-affirmation crutch) - "We're not just building software, we're changing lives" - "In today's fast-paced world..." - "Let's dive in..." / "Let's unpack this..." - Excessive bullet points where prose would flow better - Bolding every other phrase - Meaningless transition sentences - Headers every 2 paragraphs in casual content 6. **Never assume the reader shares your context.** (Curse of Knowledge) 7. **Never tell an unbelievable story without first building the world that makes it believable.** If your truth sounds like fiction, you haven't done enough world-building. 8. **Never skip grounding.** Every claim without an example, case study, or prototype is a claim floating in air. 9. **Never make all points equal.** Hierarchy matters. Lead with the strongest, most world-fitting argument. The rest are supporting cast. 10. **Never forget: you're not writing words. You're running a program in someone else's brain.** --- ## PHASE 5: THE EXECUTION CHECKLIST Before publishing/sending ANY piece of writing, verify: - [ ] **Audience model is explicit.** You can describe their world, language, desires, and fears in specific terms. - [ ] **Opens inside their world.** First sentence would make them nod, not furrow their brow. - [ ] **Atomic units established.** Common ground is laid before ANY new idea is introduced. - [ ] **World-building is consistent.** No element contradicts another. - [ ] **Grounded with examples.** Every significant claim has a case study, story, or prototype. - [ ] **Single leverage point identified.** The core argument is clear and singular, not scattered. - [ ] **Cognitive load is minimized.** Simple words. Clean sentences. No unnecessary complexity. - [ ] **In-group language used correctly.** You speak their dialect, not yours. - [ ] **Stories over statistics.** Data is wrapped in narrative, not presented raw. - [ ] **Immune system respected.** New ideas are introduced gradually, after agreement is established. - [ ] **Resolution provided.** If you created tension or dissonance, you resolved it with your idea. - [ ] **Sounds human.** Read it aloud. Does it sound like a real person talking, or like a corporate-AI hybrid? If the latter, rewrite. --- ## APPLICATION TO AI PROMPTING Everything above applies DIRECTLY to prompting LLMs. This is not analogy — it's technically accurate. - An LLM without a prompt exists as pure potential — a vast probability space. - Your prompt is the physical laws of a small universe. You're constraining what's possible. - A vague prompt ("tell me a story") gives the model no direction. It wanders to the most generic, statistically likely output. - A detailed prompt ("write a detective story set in 2049 cyberpunk India where the protagonist communicates in gang signs") teleports the model to a highly specific region of its embedding space. **You are spawning universes every time you prompt.** A three-line prompt makes a generic, shapeless universe. A detailed world-building prompt — with examples, constraints, anti-patterns, tone, audience, and micro-details — creates a rich, specific, alive universe. Think of it as a clay pot. Your words shape it. If you only press on one side, the rest stays generic. The more sides you shape — the more laws, lore, examples, and constraints you define — the more specific and useful the output. Advanced prompting IS world-building. Few-shot examples are case studies that root your world. Negative examples ("don't do X") are physical laws. Persona definitions are character backstories. Every principle in this document works on AI exactly as it works on humans. --- ## QUICK-REFERENCE: THE PERSUASION STACK ``` Layer 0: KNOW THE MIND → Who are they? What world do they live in? Layer 1: ENTER THEIR WORLD → Use their language, validate their beliefs Layer 2: BUILD COMMON GROUND → Atomic units, shared truths, nod-along statements Layer 3: INTRODUCE THE SHIFT → Small modification to their existing world Layer 4: GROUND IT → Examples, case studies, prototypes, micro-details Layer 5: CREATE TENSION → Cognitive dissonance between two things they believe Layer 6: RESOLVE WITH YOUR IDEA → Your proposal as the natural resolution Layer 7: MAKE IT EASY → Simple language, clean structure, cognitive hospitality ``` Each layer must be solid before building the next. Skip a layer and the structure collapses. --- *Remember: The amateur writes to express. The professional writes to INSTALL — a belief, a feeling, a decision — in someone else's mind. Your writing is not about you. It never was.*