This directory contains all social media content for the Redistributed publication.
Directory Structure
social/
├── README.md ← You are here
├── queue/ ← Scheduled posts and queue state
│ └── schedule.json ← Publishing queue (managed by queue-manager.js)
├── threads/ ← Multi-post thread content (evergreen)
├── reactive/ ← Quick-reaction posts to breaking news
├── responses/ ← Engagement replies and discussion starters
└── templates/ ← Reusable templates and platform guides
Content Types
Reactive Posts (reactive/)
Time-sensitive responses to AI news. These should be drafted within hours of a story breaking and connect the news back to Redistributed's redistribution analysis.
Format: Short (200-400 words), punchy, opinionated through redistribution lens. Turnaround: Same day as news breaks. Platforms: All (adapted per platform by post-formatter.js).
Thread Explainers (threads/)
Evergreen educational threads that break down complex topics into digestible multi-post sequences. These can be posted at any time and reposted periodically.
Format: 8-12 numbered posts, each under 280 characters. Includes LinkedIn consolidation. Turnaround: 1-3 days for research and drafting. Platforms: Twitter, Bluesky, Mastodon (threads), LinkedIn (consolidated).
Engagement Responses (responses/)
Replies and discussion starters for engaging with trending conversations. Includes templates for common debate patterns and platform-specific response strategies.
Format: Tweet-length responses with variations for different tones. Turnaround: Real-time during active engagement. Platforms: All, with platform-specific adaptations.
Templates (templates/)
Reusable frameworks including:
- Response templates for 10 common AI debate patterns
- Platform-specific engagement guides
- Community questions and polls
- Quote-tweet/repost templates
Frontmatter Schema
All social content files use this YAML frontmatter:
---
title: "Post title"
type: reactive | thread | engagement | cross-promo | question | data-snapshot
platforms: [twitter, bluesky, mastodon, linkedin, reddit]
topic: "AI topic category"
evergreen_links: ["slug-of-related-article"]
status: draft | ready | scheduled | posted
scheduled_date: 2026-03-15
tags: [labor, policy, concentration]
priority: high | medium | low
news_url: "https://..." # for reactive posts
created: 2026-03-12
---
Status Flow
draft → ready → scheduled → posted
- draft: Being written or needs revision
- ready: Approved for posting, awaiting scheduling
- scheduled: In the queue with a date/time
- posted: Published (kept for archive/reference)
Tooling
All social workflow scripts live in /scripts/social/:
| Script | Command | Purpose |
|---|---|---|
topic-scanner.js | pnpm social:scan | Scan RSS feeds for trending AI topics, score for redistribution relevance |
post-formatter.js | pnpm social:format | Format content for specific platforms (character limits, threading) |
evergreen-linker.js | pnpm social:link | Match trending topics to existing Redistributed articles |
queue-manager.js | pnpm social:queue | Manage publishing schedule, track what's posted where |
Daily Workflow
- Morning scan:
pnpm social:scan— check for overnight AI news worth reacting to - Triage: Review scanned topics, create reactive posts for high-relevance items
- Link:
pnpm social:link --scan-queue— find evergreen article connections - Format:
pnpm social:format --format-all --preview— preview platform-specific versions - Schedule:
pnpm social:queue --add <file>— add to publishing queue - Post:
pnpm social:queue --next— see what to post, then post manually - Engage: Monitor responses, use templates from
templates/to join discussions - Archive:
pnpm social:queue --mark-posted <id> --platform <platform>
Weekly Review
pnpm social:queue --stats— review posting frequency and distributionpnpm social:queue --gaps— identify scheduling gapspnpm social:link --suggest-gaps— find content gaps (trending topics with no articles)- Review engagement metrics and adjust strategy
Voice Reminders
When writing social content, remember:
- Lead with redistribution: Who benefits? Who pays? Where does value flow?
- Be specific: Name companies, cite numbers, reference policies
- Show the mechanism: Don't just say "AI is unequal" — show how the inequality works
- Respect intelligence: No dumbing down, but no jargon gatekeeping either
- Engage honestly: Acknowledge uncertainty and genuine disagreements
- Name your sources: Reference specific thinkers (Crawford, Winner, Eubanks, Acemoglu, etc.)