linkedin-post-writer

name

linkedin-post-writer

description

Draft LinkedIn posts in Shaw's personal writing style. Use this skill whenever Shaw asks to write, draft, brainstorm, or edit a LinkedIn post — including phrases like "write a post about", "help me draft a LinkedIn update", "turn this into a LinkedIn post", "post idea", or any reference to writing for LinkedIn. Also trigger when Shaw shares a topic, milestone, or idea and asks to "make it a post" or "write this up." Even casual requests like "can you LinkedIn-ify this" or "help me announce X" should trigger this skill.

LinkedIn Post Writer

Write LinkedIn posts that match Shaw's voice and content philosophy. The examples are the primary guide — read them, internalize the feel, and let them shape your drafts more than any rule in this document.

How to Use Examples

Examples are split by post type in the references/ folder:

• examples-educational.md — Builder's Logs, frameworks, tech stacks, how-I-do-X
• examples-thesis.md — Idea-driven opinion posts, trend analysis, bigger ideas
• examples-personal.md — Life milestones, travel, birthdays, first wins (Shaw's highest-performing archetype)
• examples-video-bridge.md — Posts that bridge to a YouTube video or blog
• examples-promo.md — Event promos, resource promos, milestone announcements
• examples-event-promo.md — Promos for Shaw's own live events: the proof-dense insider-access shape

After choosing a post direction, read the matching example file(s). If the post bridges to a video, also read examples-video-bridge.md. Pick 2-3 specific examples as your reference points while drafting.

After drafting, compare your draft to those examples. Does it feel like them in density, weight, and tone? If the examples wouldn't include a section you wrote, that section is probably scaffolding. Cut it.

Hard Rules

These are guardrails, not construction instructions. They catch mistakes — they don't tell you what to build.

• Em dash limit. At most one em dash ("—") per post. Prefer zero.
• No clichés or hype. No metaphors, idioms, or business jargon. Forbidden: "game-changer," "move the needle," "raise the bar," "secret sauce," "level up," "hit the ground running," "low-hanging fruit," "unlock" (figurative), "double down," "win-win," "take it to the next level," "circle back," "think outside the box."
• No motivational CTAs. Forbidden: "What inspires you?", "What's your best lesson?", "What motivates you?", "Never give up", "It compounds."
• No tricolons with negation. Forbidden pattern: "No X, no Y—just Z."
• Authenticity. Never fabricate Shaw's experiences, results, or opinions. Only reference what Shaw provides, what appears in past posts, or established facts.
• No unverifiable scale claims. No "infinite," "limitless," "no ceiling." Bounded ranges ("10x, 100x") are fine.

Writing Principles

These explain why Shaw's writing works. Internalize them rather than treating them as a checklist to comply with.

The examples are the source of truth. When a principle and an example seem to conflict, follow the example. The examples show what Shaw actually publishes. The principles are approximations.

Brain dumps are raw material, not a blueprint. When Shaw talks through an idea — in chat, voice, or a sprawling Notion note — he's surfacing every thought at once: hedges, restarts, and the same point made three different ways. This is ore to mine, not a structure to transcribe. Pull out the genuine insight, the specific nuggets, and the nuance, then set the transcript aside and build the narrative from scratch. Never carry the brain dump's order or in-the-moment phrasing into the post. The tell that you've done this wrong: your draft echoes the sequence and wording of what Shaw just said. When that happens, the post reads like a cleaned-up transcript instead of a composed piece — and Shaw will end up rewriting it into a tighter arc you could have found yourself.

The thesis is usually buried, not stated first. In a brain dump the real payoff often arrives last, mid-thought, or as the "I guess the nuance is…" aside — rarely in the opening sentence. Find the load-bearing insight wherever it sits and let it anchor the arc, rather than leading with whatever Shaw happened to say first. The opening line he reaches for while thinking out loud is almost never the hook the post wants.

Show, don't scaffold. If the content demonstrates a concept, don't define it first. A 4-step process that creates Skills shows what Skills are — a preceding paragraph explaining "what are Skills" is redundant. Before including any setup or definition, ask: would the post still make sense without this? If yes, cut it.

Lean by default. Every element — descriptions, closers, conceptual bridges, jargon definitions — must earn its place. If the reader could infer it, cut it. List items that are self-explanatory don't need descriptions. A body that naturally leads into the video bridge doesn't need a separate closer. Shaw's best posts feel like every word is load-bearing.

Frameworks earn expansion when they're the payoff. If a post references a framework as a side note or callback, naming it is enough: "a 4-step roadmap (Equip → Educate → Adopt → Redesign)." If a framework is the post's payoff, each item earns its own line with a one-line definition using "= " format: "Step 2: Educate = show people what they can do with the tools and how it's relevant to their work." Ask: if the reader never clicks away, does the framework land? Self-contained posts need the expansion; bridge posts can get away with the name.

Generous peer, not thought leader. Shaw shares what he's found, in case it's useful. The energy is "here's what I found" not "here's what you should know." Light self-deprecation (😅, :P, 😬) undercuts anything that might read as bragging — but only in posts with an emotional register (milestones, personal stories). Educational and promotional posts don't need it; inserting personal asides can dilute the focus.

Specificity is credibility. Concrete numbers, tool names, dollar amounts, real outcomes. Never generic observations anyone could make. Don't write "I see this in my work and in the teams I train." Write "Claude Code writes ~99.9% of my code."

One idea, one post. Never cover two topics. If you can't state the post's single idea in one sentence, it's too broad. This applies to the solution too: keep it as single as the problem. If the payoff is "stop consuming, start building," don't expand it into a list of example tactics ("automate the task you hate, build the tool you wish existed, fix the thing that's been bugging you"). One clean move lands harder than three illustrations of it.

Validate the hard part before offering structure. When a thesis asks the reader to do something difficult, acknowledge the difficulty before presenting the solution. A line like "this won't be as easy as you'd like, but it is a simple roadmap" disarms the "easy for you to say" objection that otherwise makes clean frameworks read as glib. Especially useful in thesis posts that introduce a roadmap or model — without the validation beat, the reader files the post under "consultant energy" and scrolls past. But validation doesn't always need its own line. When a post opens by naming the reader's situation precisely (the exact fears, the exact thought they're having), that accurate naming is the validation, and a separate "I get it, this is hard" beat on top of it reads as padding. Reserve the explicit validation line for when you're about to hand the reader a difficult ask or framework, not as a reflex.

Technology is personal. Never discuss tech in the abstract. Ground it in Shaw's specific experience — what he pays, what he uses, what broke, what clicked.

Parentheticals add texture. Shaw uses parenthetical asides naturally — "(and it is)", "(which is the fun part)", "(huge)". They add qualifiers and real-time context without breaking flow. Don't force them in, but notice that Shaw's posts usually have a few.

Reframe with contrast, not explanation. When the post turns on a reframe, build it as two mirrored halves rather than a sentence that explains the mechanism. Shaw's strongest version of "people don't act because the payoff is uncertain" is "the cost is clear (using a new tool is immediately painful) while the benefit is not (how do I know this will pay off?)" — two parallel clauses, each carrying its own parenthetical example. The symmetry itself does the persuading; an explanatory "they can't answer this because…" sentence is weaker. When you catch yourself explaining why something is true, check whether a clean A-vs-B contrast would land harder.

Put the reader inside the decision. At the emotional or logical crux, drop into first person so the reader runs the thought themselves instead of watching a third party. "If no one on my team uses Claude, I probably won't either. But if someone is getting real results, I'll give it a proper shot" beats the third-person "if no one on the team uses it, they won't either." First-person interior monologue at the turn makes the reader feel the heuristic rather than observe it. Reach for it whenever the post hinges on what someone is thinking or feeling.

Lead with stance, bury the ask. Shaw's highest-performing posts open with an opinion, insight, or personal experience — not a promotion. When a post has a CTA (event, resource, link), the ask comes in the final 20% after the reader already got value. Posts that lead with a CTA or "I'm hosting X" consistently underperform posts that lead with a thesis and weave the CTA in at the end.

Write for the repost. Reposts are Shaw's primary reach amplifier — they account for 26-65% of all engagement across 9 months of data. Posts that get reshared tend to be: (1) self-contained insights someone can endorse by sharing, (2) frameworks or lists that make the sharer look informed, (3) opinions that let the sharer signal alignment. When drafting, ask: "Would someone repost this to say 'this is what I think too'?" If yes, you're on the right track.

Every line moves forward. If a line restates the hook or echoes an earlier point, cut it. The reader should learn something new from every sentence. The most common offenders: throat-clearing transitions ("here's what I've noticed," "but here's the thing," "here's the problem"), restating the hook in different words a few lines down, and anaphora triples ("Another model... Another headline... Another reminder...") that feel rhetorical but carry no new information. Cut the connective tissue and let the lines sit next to each other. A single flat declarative ("Everything else is noise.") can carry a turn that a multi-line build would only dilute.

But don't over-compress into clinical. This principle targets lines that carry no new information, not the connective warmth that makes Shaw sound like a person thinking out loud. A short rhetorical question that genuinely opens the next move ("So what does it take to drive behavior change?") and the occasional light transition are part of his voice, not fat. The test is whether the line advances or just echoes — a question that sets up new substance advances; "here's the thing" before restating the hook doesn't. When in doubt, Shaw's real drafts run a touch warmer and looser than maximal compression would produce; match that register rather than stripping every connective.

One flat declarative, not a stack. A single flat declarative can carry a turn — but resist manufacturing mic-drops. Stacking two or three aphoristic closers ("It won't happen by buying more seats. It happens one result at a time.") reads as trying too hard and undercuts the one line that mattered. When the framework or insight is the payoff, let it be the payoff and land on one quiet line, not a closing flourish. If you've written a punchy couplet to "button up" the ending, it's usually a sign to cut one half or both.

Formatting

Mobile-first LinkedIn consumption. These are sensible defaults, not mandates.

• One sentence per line. Treat line breaks like punctuation.
• Numbered lists for steps and ranked items. Keep list items short — cut sub-clauses and qualifiers into prose before or after the list.
• Bullets sparingly. For quick lists within a narrative, never as the entire post structure.
• Emoji as functional punctuation. 👇 = scroll down, 👉 = CTA/link, 😅 = self-deprecation, 🥳 = celebration. Max 2-3 per post. Never decorative.
• One link in the body. Extras go in comments after posting.
• Footer CTAs are optional. If using, separate with -- and keep to 1-2 lines.

Post Types

These are loose shapes, not rigid templates. The examples show how they actually play out — read the relevant example file to see the shape in practice.

Builder's Log — Process, tech stack, how-I-do-X. Walk through tools or steps. Practical, no fluff. Can carry 6-8 short items. (See examples-educational.md)

Thesis Post — Opinion, trend, bigger idea. Build a narrative arc: status quo → tension → resolution. 2-3 conceptual moves max. Include a "so what" connecting to broader stakes. When the thesis corrects a belief, reframe in one line rather than explaining the mechanism behind it: assert that the thinking is wrong and give the truer frame, then move on. Don't lecture on why the reader feels that way. A reliable micro-structure for emotion-driven theses: name the reader's situation precisely → quote the giving-up thought → one-line pivot ("but that's more anxiety speaking than reality") → the reframe → a flat-declarative punch → a directive close. (See examples-thesis.md)

Personal / Life Milestone — Shaw's single highest-performing archetype. A concrete life event (travel, birthday, career transition, first win) becomes the anchor for one earned insight. Numbers do the credibility work, light self-deprecation protects against a brag read. Use when Shaw shares something that actually happened to him personally. (See examples-personal.md)

Framework Post — Educational breakdown of a model or mental framework. Define levels, stages, or categories. Each one earns its place with a concrete example or one-line explanation. (See examples-educational.md)

Milestone/Announcement — Where you were → what was hard → what you learned. Anchor to specific numbers and timelines. (See examples-promo.md)

Video/Blog Bridge — The post makes people care about the topic (the why). The video teaches it (the how). Don't mirror the tutorial's structure — that makes the post redundant. The post should stand alone as valuable even if nobody clicks. Frame the bridge naturally: "In a recent YouTube video, I walk through the full framework with concrete examples." (See examples-video-bridge.md)

Promo/Resource — Event promos, free resource shares. Value-first framing: what the reader gets, not what Shaw is selling. Soft CTAs woven into narrative, never the main event. The highest-performing promos are disguised as thesis or milestone posts — they open with an insight or story and only reveal the CTA in the last few lines. Pure promo posts (CTA in hook or first half) get 3-5x less reach. When possible, structure promos as a thesis post with a P.S. link rather than an announcement with a value add. (See examples-promo.md) For Shaw's own live events (webinars, workshops), read examples-event-promo.md instead — those carry a direct CTA and follow the proof-dense insider-access shape.

Conversion handoff: If the explicit goal is conversion — selling spots, driving paid signups, not just awareness — use conversion-copy's channel-linkedin-posts.md. It has the 3-question filter and conversion-specific hook tactics. Keep using this skill for general voice and post-shape patterns regardless.

Writing Process

Step 1: Read examples

Read the matching example file(s) from references/ based on the post type that best fits Shaw's content. Pick 2-3 specific examples as reference points for this draft.

Step 2: Clarify purpose

State in one line each:

• Purpose: What is this post communicating?
• Audience: Who benefits?
• Post type: Which shape?
• Reference examples: Which 2-3 examples are guiding this draft?

Step 3: Find the narrative, then write the body

If Shaw handed you a brain dump (chat, voice, or a loose Notion note), don't start writing yet. First separate the raw material from the narrative. List the actual ideas, nuggets, and nuance buried in what he said — then identify which one is the load-bearing insight (it's usually not the thing he said first; see "The thesis is usually buried"). Sketch 2-3 different narrative arcs that could carry that insight, note what each leads with and where it turns, and pick the one that lands hardest. The first arc that comes to mind is almost always just the brain dump's order in disguise, so generate alternatives before committing.

Then write the body. This is the substance. Ground every claim in Shaw's actual experience. Build the arc you chose — do not transcribe the brain dump into post form. Let the reference examples guide density and weight; if your body is twice as long or twice as dense as the examples, something is off.

Step 4: Brainstorm hooks

This is the most important step. Spend real effort here.

If the source is a Notion page, the page title is Shaw's intended hook direction. The briefing/body is the substance; the title is the angle he's already leaning toward. Anchor your hook options to the title's framing first, and only diverge if Shaw asks. Example: a page titled "Why AI tools (alone) are not enough" should produce hooks built on that claim, not on the post's eventual payoff (the system of record).

Write 3-5 hook directions. For each, write why it could work and what risk it carries. Consider the hook's promise — "here's why" promises reasoning (Thesis Post), "here's how" promises process (Builder's Log). The hook determines the post structure.

Then pick the strongest and write the full 1-2 line hook.

Lead with the moment, not the concept. Shaw's strongest hooks open with something that happened — a milestone, a number he hit, a decision he made, a line a client said — not with the idea or lesson the post is about. When turning an abstract insight into a hook (especially client or strategy material), find the concrete moment that contains it and open there; the lesson is the payoff, not the hook. Idea-led openers ("Lock your AI down too hard and nobody uses it") consistently underperform moment-led ones ("A VC CFO told me asking Claude about its own security settings scared her even more").

Hook archetypes that work for Shaw:

• Time-anchored milestone: a timestamp plus a personal stake — "Last week, I crossed 75,000 subs," "15 months ago, I launched X. Today, I'm closing it down," "30 days ago, I set a goal. Today, the result." This is his strongest archetype, and the only one that reliably wins both reach and engagement rate: it's specific, true, and opens a before/after loop.
• Contrarian claim: "Here is something I believe (that people will disagree with)."
• "Most people get X wrong": name what the majority does, then position as the corrector — "Most founders guess their ICP…", "Most developers use Claude Code like a chatbot."
• Bold framing / flat thesis: a certain, unhedged claim, often followed by a question — "The cost of generating code is going to 0. The question is, what replaces it?"
• Challenge/countdown: "I have 29 days to ship an AI app."
• Problem/solution: Start with the problem the reader faces.
• Specific number: dollar amounts, time periods, or ranked lists in the opening line — "Why I pay $160/mo for AI tools (and how I use them)", "3 side hustles that made me $23,603.96", "10 YouTube channels that keep me ahead." Use exact figures, never rounded ($23,603.96 reads true; "$23k" reads like marketing). Concrete numbers consistently outperform abstract claims.

Hooks are two-line systems. The second line pulls the reader forward — ellipsis, "And", "But", or playful punctuation (:P, 😅) can all create momentum past the "see more" fold. The shortest viable hook usually wins ("Back to building…" landed at 2.67%). There's no single right technique. Study how the reference examples open and match the energy.

Step 5: Write the CTA

Usually one clear, specific question. No "or" / "and" splitting attention. Answerable in one sentence but invites real thought. But a question isn't mandatory. When the thesis is itself a call to action, a directive close ("stop consuming content about AI and start building with it") often lands harder than a discussion question tacked on. Match the ending to the post; don't force a question where an imperative is the natural close.

Prefer outcome questions over activity questions. "What AI skill resulted in the biggest transformation for your work?" outperforms "What AI skill are you developing right now?" because it invites concrete stories and results rather than status updates. Past-tense, result-focused CTAs generate testimonial-style replies that add credibility for later readers and are more shareable than in-progress updates. When in doubt, ask about an outcome someone has already experienced, not an activity they're currently doing.

Step 6: Assemble and compare

Put together hook + body + CTA. Then compare the full draft to your reference examples:

• Does it feel like the same person wrote both?
• Is it the same density and weight?
• Would the example posts include every section you wrote?

If something feels like scaffolding the examples wouldn't have, cut it.

Step 7: Present the draft

Show Shaw the completed post. Note which examples guided the draft.

Performance Insights

Data patterns from Shaw's LinkedIn analytics (Q3 2025 – Q1 2026 dashboard pulls, plus a full-census feed-API pull of all 274 original posts through June 2026). Use these to inform drafting decisions, not as rigid rules.

Impression trends. Shaw's median impressions per post grew from ~4,500 (Q3 2025) to ~8,200 (Q1 2026). Top posts consistently hit 15K-30K+. The growth correlates with more thesis/opinion posts and fewer pure promos. (Note: that ~8,200 reflects the dashboard's top ~50 posts by reach. The full-census median across every post is lower, ~1,320 — most posts are modest and a few break out.)

Reach and engagement rate are different games — pick one per post before writing the hook. In the full census (median ~1,320 impressions, median 1.63% engagement rate), the highest-reach hooks are lists, money numbers, and life events, but they pull a low rate (a 52K-impression list post at 0.87%, a $23,603.96 side-hustle post at 0.21%). The highest-rate hooks are personal, staked, and milestone-driven (ICP-with-AI at 4.15%, "ship 1 product/month" at 3.94%, the 75K-subs milestone at 3.55%). One hook rarely wins both. Decide whether a given post is for reach or resonance, then choose the matching hook type.

What drives reach vs. what drives rate (June 2026 28-day pull: 23 posts, top-half rate 2.07% vs bottom-half 0.92%). The two games decoupled again on fresh data — the single widest post (5,689 impressions) ranked near the bottom on rate (0.97%), while the highest-rate post (3.16%) had the smallest audience (316). Reach and resonance are not the same axis, and the widest post is usually not the one that resonates.

Patterns that earn reach:

• Opportunity / recruiting / open-spots posts. These spread through tags, shares, and reposts into networks beyond Shaw's followers, so they pull the most impressions by far — but the borrowed audience is low-intent, so the rate craters. Reach mechanism, not a resonance pattern; use deliberately when the goal is volume.
• Contrarian or opinion stance (both "contrarian take of the week" posts landed in the top four by reach). A debatable claim earns the early dwell and comments that trigger distribution.
• A specific number in the hook ("5,000 to 400," "10 hr/wk," "8 Skills").
• Broad, relatable framing — jobs, the AI skill gap, how AI feels — over narrow tool tactics.

Patterns that earn rate:

• Personal / community / host posts to a tighter, invested audience (the DFW dinner at 3.16%, the AgentCon session at 2.28%). Less reach, more resonance.
• Behind-the-scenes reveals ("the 8 Skills that run my business," 2.15%).
• Identity / opinion lines the reader wants to co-sign ("I want my AI better than me," 1.72%).

The rare both-winner is a sharp contrarian claim — "the gap isn't tools, it's a skill" hit 1,162 impressions at 2.93%, top quartile on both axes. When a post must do double duty, that's the shape to reach for.

Reposts dominate engagement. Across all periods, reposts account for 26-65% of total engagement. Comments and reactions are secondary. This means reach is driven more by shareability than by comment-bait CTAs.

Post type performance. Thesis and opinion posts consistently outperform other types on impressions (2-5x vs. pure promos). Builder's Logs and framework posts perform well when they include a specific number or bold claim in the hook. Video bridge posts perform moderately — the post needs to stand alone as valuable.

What underperforms. CTA-first promos, generic "here's what I learned" framing without specifics, and posts that open with announcements rather than insights. Also weak on rate: narrow product-tip posts ("5 Cowork features 95% miss" at 0.52%, "use AI better than 95% of businesses" at 0.78% — niche tool tactics reach a small, already-converted audience and pull a low rate), recurring "this question keeps coming up in my workshops" openers, and a second same-day post that cannibalizes the first (the post published behind the AgentCon announcement bottomed out at 0.22%).

Post Length

Most posts: 150-300 words. Listicles can run longer. Never pad for length.

Thesis posts skew shorter than you'd think. A thesis built on a single reframe often lands best at ~120-150 words: long enough to name the tension and turn it, short enough that every line is load-bearing. Treat 400 as a rare ceiling for genuinely multi-move arguments, not a target. When in doubt, cut. The tighter draft almost always hits harder than the fuller one.