newsletter-writer

name

newsletter-writer

description

Draft newsletters for Shaw's Kit list (The Data Entrepreneurs) in his exact voice and structure. Use this skill whenever Shaw asks to write, draft, edit, or repurpose a newsletter — including phrases like "write a newsletter", "draft a Kit broadcast", "turn this post into a newsletter", "weekly send", "newsletter idea", "make this a newsletter", "the next email to my list", or any reference to writing for his email list. Also trigger when Shaw shares a topic, milestone, framework, observation, or LinkedIn post and asks to "make it a newsletter" or "send it to the list." Even casual phrasing like "what should I send this week" or "I want to write something to my subscribers" should trigger this skill. Covers both content/teaching newsletters and sell newsletters — for deep conversion-focused diagnosis on a sell email, hand off to `conversion-copy` after the draft is in place.

Newsletter Writer

Shaw's canonical skill for drafting newsletters to his Kit list (~1.3k active subscribers — pruned from a larger list by removing the "30 AI Projects" guide cohort, who were mostly technical/engineering and not the founder-buyer ICP). The active list now skews toward founders, operators, and AI-curious professionals aligned with ABA's target customer.

The skill is built on patterns extracted from his sent archive — the moves that show up in nearly every email he ships, in the rhythm and length that consistently performs.

Before drafting

Lock these:

• Mode: content/teaching, observation, or sell? (Determines the template and the P.S.)
• One core idea: every newsletter has exactly one. If there are two, split it into two emails.
• Source material: post being repurposed, framework being taught, observation being shared, or offer being sold. If Shaw hands over a LinkedIn post, draft, or rough notes, lean on it heavily — his existing language is usually closer to the final voice than anything you'd invent.
• Length target: 250–450 words. If you're approaching 500, cut.

If any of the above is unclear, ask before drafting.

Voice (non-negotiable)

Shaw's voice is a confident peer-to-peer operator. Smart, no-fluff, declarative. Not coaching, not hype, not corporate.

• Cold open. Never "Hey," "Hi friends," or any greeting. Open with the claim, the scene, or the problem.
• One thought per line. Short paragraphs, often single sentences. Heavy whitespace. The page reads like a stack of statements, not a wall of prose.
• Declarative over coaching. State the thing and move on. No "that's fine," "don't worry," or pep-talk reassurance.
• Anti-hype. Kill landing-page words: unleash, supercharge, transform, empower, leverage, revolutionary, game-changing, cutting-edge, next-gen, seamless, robust, elevate, exceed.
• Visceral verbs. "Delete 20 hours" beats "save 20 hours." "Hand off" beats "automate." "Kill the bottleneck" beats "reduce it."
• Specific over vague. Numbers, named roles, concrete examples. "165 alumni, 4.7-star rating" beats "lots of happy students." "The head of data at an enterprise" beats "some leaders."
• Light on em-dashes. Shaw has asked to ease off them. Use commas, periods, or parentheses. One or two per email is fine.
• Confident first person. Shaw uses "I" and "you" generously. He doesn't hide behind "we" or passive voice.

Subject lines

Three patterns that perform in his archive. Pick whichever fits the mode.

Curiosity gap (ellipsis) — best for personal/announcement or contrarian takes. His top-performing subject style.

• "Something's changed..." (43% open)
• "Why AI transformations fail..." (27% open)
• "You're Not Behind: What's (not) new in AI"

"How I / How to" — best for content/teaching emails.

• "How I Turn AI Ideas into Applications"
• "How to learn AI skills (that won't go obsolete)"
• "How I delegated 15 hrs/week to Claude"

Concrete contrast — best for sell or strong-result emails.

• "3 weeks of waiting, done in 1 hour"
• "If you only learn one AI skill, learn this."

Avoid emojis in subject lines (rare in his archive — they belong in the P.S. or body if anywhere). Avoid all-caps, exclamation marks, urgency theater.

When uncertain between two subject lines, offer both with reasoning rather than pre-selecting.

Kit tool caveat (subject + preview)

The Kit get_broadcast API doesn't mirror what's actually in the Kit editor: its preview_text field goes stale (it returned a leftover placeholder that didn't match the real preview), and its subject field exposes only a single subject — an A/B subject test is invisible to the API. So when capturing a broadcast (e.g. onto its ABN page) or reasoning about what shipped, treat the subject lines and preview text Shaw provides as the source of truth, not the API fields. The broadcast body/content is reliable; the subject and preview_text metadata are not.

The two templates

1. Framework email (most common)

Used when teaching a process, set of lessons, or steps. ~70% of his content sends.

Structure:

1. Cold open — 1–3 short lines. Sets up the problem, the personal credibility, or the contrarian observation.
2. Reframe / principle — the "not X, but Y" pivot. Often called out with └→ or →.
3. Numbered steps or lessons — usually 3–5. Each one: title + 1–3 short paragraphs.
4. Close — short summary line, optional question hook for replies.
5. Signoff — "Keep building, / Shaw"
6. P.S. — single CTA (free resource for content emails, paid offer for sell emails).

Example openers (study the rhythm):

• "Over the past 18 months, I shipped my first two AI SaaS products. Not because I'm some cracked, 10X developer, but because I used AI coding tools (in the right way)."
• "I've quietly handed off about 15 hours of busywork a week to Claude."
• "Although building software is easier than ever, most AI projects still stall halfway through."

2. Observation email

Used when sharing a trend, signal, or pattern from talking to people in the field. ~25% of his content sends.

Structure:

1. Cold open — names the trend or contrast.
2. Setup — "I've spent the past two weeks talking to..." or similar grounding line.
3. Named-source examples — usually 3 short paragraphs, each with a specific role + a specific thing they said or did.
4. Implication — what this means, often with a triplet of short parallel sentences.
5. Question hook (optional) — "What's one thing that has transformed for you?"
6. Signoff — "Keep building, / Shaw"
7. P.S. — relevant free resource or event.

Example named-source proof:

• "The head of data at an enterprise said business stakeholders are now vibe-coding their own dashboards..."
• "A freelance analytics engineer built an agent skill teaching Claude how to interact directly with raw data sources..."
• "A data analyst at a mid-sized company used Claude to optimize an ML model..."

The roles and contexts are specific. The proof is named, not abstract.

Signature moves (use at least 2-3 per email)

These show up across nearly every newsletter in his archive. They're not optional flavor — they ARE the voice.

1. The reframe ("Not X, but Y" / "From X → Y") The single highest-leverage move in his writing. Almost every email contains at least one.

• "AI transformation isn't about bolting a new tool onto how you already work. It's about redesigning the work around what AI unlocks."
• "Last year, this would have taken hours of building custom AI automations. Today, you can do it all without a single line of code."
• "It went from 'how do I write this code?' → 'how do I get AI to write this code?'"
• "value is generated, not by creating software, but by solving problems"
• "instead of trying to predict an unpredictable future, reflect on the certain past"

2. Visual notation (→, └→, >>) Used as deliberate signals, not punctuation.

• └→ for the principle/insight callout — "└→ If you want to know what will last, look to what has lasted."
• → for transitions inside a sentence — "Equip → Educate → Adopt → Redesign"
• >> for sequence/flow diagrams — ">> User provides input → App does something → App produces output"

3. Self-deprecating credibility Acknowledge what the reader might be thinking, then redirect. Lets you make a strong claim without sounding boastful.

• "Not because I'm some cracked, 10X developer, but because I used AI coding tools (in the right way)."

4. "Pro Tip:" parenthetical asides Micro-callouts mid-body. Adds value without breaking flow.

• "Pro tip: AI champions inside the team are a great way to foster this."
• "Pro Tip: The number of GitHub stars is a quick proxy for library quality."

5. Parenthetical emphasis Drop modifiers in parens to underline a word. Conversational, high-status.

• "working (very) hard to make a modest impact"
• "(super) early bird"
• "(not) new in AI"

6. Triplet rhythm Three short parallel sentences as a beat. Use at the close or to land an implication.

• "Create more dashboards. Validate business metrics faster. Build better analytical models."
• "No code. No new tools. No engineering background."

7. Specific numbers and named sources Always anchor claims. "165 alumni" beats "many alumni." "The head of data at an enterprise" beats "a leader."

Close patterns

Standard signoff (near-universal):

Keep building,
Shaw

The new "How I delegated 15 hrs/week" used "-Shaw" instead — that's a departure. Default to "Keep building, / Shaw" unless Shaw explicitly asks otherwise.

Optional question hook before signoff when you want to drive replies:

• "Which one of these lessons resonated with you? Hit reply and let me know :)"
• "What's one thing that has transformed in your analytics work because of AI?"

Question hooks are cheap engagement — they warm the list without asking for a click.

P.S. patterns (always present, always a CTA)

The P.S. is never empty. It's always a single CTA. Three flavors based on email mode:

Free resource (default for content/teaching emails):

• Link to a YouTube video that goes deeper
• Link to a free workshop or live event
• Link to a free course or guide

Soft offer (when nudging toward a paid product):

• Lead magnet that bridges to the paid offer
• Discount code with light scarcity

Hard sell (only for dedicated sell emails):

• Direct pitch with price, scarcity, and CTA link
• Used sparingly — most P.S. lines are free resources

The P.S. should restate or extend the value of the email, not introduce a totally new topic.

Length

• Content/teaching: 250–450 words.
• Observation: 200–350 words.
• Sell: 300–500 words.
• Personal/announcement: can run longer (up to 600), but only when the personal narrative warrants it.

If the draft runs past 500 words on a content email, cut. Compression is a feature.

Output 2-3 versions, not one

First-pass copy is rarely the best. When drafting:

• Offer 2–3 angle variants for the opener specifically (e.g., one leads with the problem, one with personal credibility, one with the contrarian reframe). State which you'd pick and why.
• For the rest of the email, one strong draft is fine — but call out alternate phrasings for any line that's load-bearing (the reframe, the close, the P.S.).
• Don't pre-select silently. Show your work so Shaw can iterate.

Things to avoid

• "Hey friends," / "Hi everyone," — never. Cold open or nothing.
• "Best," / "Warm regards," — never. "Keep building, / Shaw" or similar.
• "Unleash," "supercharge," "transform" — landing-page words.
• Multiple offers in one email — never. One CTA per send.
• Empty P.S. — wasted real estate.
• Overusing em-dashes — Shaw has explicitly asked to ease off.
• Generic claims without numbers or named sources.
• Coaching tone ("don't worry," "you've got this") — Shaw doesn't reassure, he states.
• Walls of text — break up with whitespace, single-sentence paragraphs.
• Subject lines with all-caps, multiple exclamation marks, or "Don't miss out!" energy.

Handoffs

This skill is the canonical newsletter skill. Lanes:

• Deep diagnosis on a sell newsletter → after drafting, hand off to conversion-copy (newsletter channel) for line-by-line filter checks (visualize / falsify / unsignable).
• Hormozi-style / Mozi Minute newsletters → still this skill. If Shaw explicitly wants the Mozi Minute structure (one idea + stories + reasons + soft CTA), pull the framework from the Hormozi Content Framework resource in Notion (https://www.notion.so/[page-id]) and apply it here. Most of Shaw's newsletters are not Mozi-style — they're frameworks or observations.
• LinkedIn posts (not newsletters) → linkedin-post-writer.
• General email drafts (not list broadcasts) → email-helper.
• Long-form sales letter for a paid offer → sales-letter-writer.
• Storing a draft on its ABN Notion page → see notion-helper ("Capturing a Kit/newsletter draft into an ABN page") for the page format.

If Shaw asks to repurpose a LinkedIn post into a newsletter, this is the right skill — the source material is a post, but the deliverable is a newsletter, and the deliverable's voice and structure dominate.

Worked example: repurposing a LinkedIn post into a Framework email

Source post (raw):

I've delegated 15 hr/wk of busywork to Claude. My approach was surprisingly simple — no coding or new tools needed. It's a 4-step process: (1) Do it once with Claude (2) Distill into skill (3) Test it (4) Improve it.

Newsletter draft (Framework template):

Subject: How I delegated 15 hrs/week to Claude Preview: Pro tip — treat Claude like an intern, not a tool.

I've quietly handed off about 15 hours of busywork a week to Claude.

This includes: email drafts, pre-call research, CRM updates, LinkedIn posts, outreach campaigns, lead qualification, and daily briefings.

Last year, this would have taken hours of building custom AI automations. Today, you can do it all without a single line of code.

The shift was treating Claude like an intern instead of a tool. In other words, thinking of the task as onboarding rather than engineering.

Here's the 4-step loop I run through.

Step 1: Do it once (with Claude) Walk Claude through the task with you in the loop. It'll take longer than doing it yourself, but this is an investment (just like training a new hire).

Step 2: Distill into Skill Tell Claude to turn what you just did into a reusable skill. It writes its own SOP so you don't have to. This is important because Claude understands Claude better than we do.

Step 3: Test it New chat. Fresh input. Run the skill. Should be 5x faster than Step 1, because Claude's running the SOP instead of figuring it out from scratch.

Step 4: Improve it Skills aren't static. When Claude misses something, or you want a new capability, just tell it. This allows you to teach Claude about your work "little and often" rather than all at once.

Next Step: Look at your calendar for the past 2 weeks and see what you're spending the most time on. Then pick 1 task and try to delegate some (or all) of it to Claude.

If you get stuck, reply to this email and I'll point you in the right direction.

Keep building, Shaw

P.S. This is the exact 4-step loop we run inside my 1:1 Claude Workshops. This is the last round before they go public. 3 spots left, first-come, first-served. Book your fit call →

What this draft did well (in this skill's terms):

• Cold open (no "Hey,").
• One thought per line.
• Compare-and-contrast pair: "Last year [X]. Today [Y]."
• Reframe: "intern instead of a tool" — names the shift cleanly.
• Numbered framework with title + 2 short lines per step.
• Pro Tip-style aside via the analogy in Step 1: "(just like training a new hire)".
• "Little and often" — aphoristic punchline.
• Engagement layer (reply if stuck) before the offer P.S.
• P.S. is a single hard-sell CTA appropriate for this mode.
• "Keep building, / Shaw" signoff.

One last thing

The newsletter is in the inbox. Brands don't belong in the inbox. Be a person.

If a draft sounds like a marketing system wrote it, rewrite it until it sounds like Shaw on a Tuesday morning, talking to one operator who already knows what he does.