# Blog Content Ideas — Living Document

**Purpose:** Central repository for blog post ideas, content direction, and pipeline input.  
**Updated:** 2026-04-21  
**Audience:** Matt (Director), Content Pipeline, The Board

---

## How to Use This Document

1. **Drop ideas** in the appropriate category (or create new ones)
2. **Tag status:** `idea` → `drafting` → `in-pipeline` → `published`
3. **Pipeline reads** this file for content direction and prioritization
4. **Markdown-native:** Works with Obsidian, Git, and automated pipelines

---

## Content Categories

### Home Lab Growing Pains
*Personal infrastructure experiments, failures, and victories*

| # | Idea | Status | Notes | Target Length |
|---|------|--------|-------|---------------|
| 1 | **The Night I Broke DNS and My Wife Couldn't Reach Facebook** | `published` | Origin story format | ~1,200 words |
| 2 | How I Accidentally DDoSed Myself with a Cron Job | `idea` | Pi hole + update script story | ~800 words |
| 3 | Building a 3-2-1 Backup Strategy on a Budget | `idea` | BorgBackup + Backblaze B2 | ~1,500 words |
| 4 | Why I Run a Dedicated NUC Just for Home Automation | `idea` | Beelink N150 justification | ~1,000 words |
| 5 | The 2 AM Incident: When Docker Networks Collapsed | `idea` | Compose networking deep-dive | ~1,200 words |
| 6 | Tailscale: The VPN That Actually Made Sense | `idea` | Zero-trust for families | ~1,000 words |
| 7 | My Router Died at the Worst Possible Moment | `idea` | Hardware failure + recovery | ~800 words |
| 8 | Why I Replaced Plex with Jellyfin (And Might Switch Back) | `idea` | Media server comparison | ~1,200 words |
| 9 | The Homelab Cost Spreadsheet: What I Actually Spend | `idea` | Transparency post | ~800 words |
| 10 | Building a "Family Cloud" That Doesn't Spy on My Kids | `idea` | Nextcloud + Photos | ~1,500 words |

### OpenClaw Hacks & Tutorials
*Multi-agent setup, configuration, and lessons learned*

| # | Idea | Status | Notes | Target Length |
|---|------|--------|-------|---------------|
| 1 | **Running Three AI Agents on a Beelink N150** | `drafting` | Socrates/Daedalus/Wadsworth origin | ~2,000 words |
| 2 | How I Debugged the "Compaction Ghost Bug" | `idea` | Cross-agent contamination post-mortem | ~1,500 words |
| 3 | Building a Content Pipeline with Local LLMs | `idea` | Gaming PC 3080 Ti setup | ~1,800 words |
| 4 | The Sovereign Constraint: Cloud Bootstrap, Local Target | `idea` | Architectural philosophy | ~1,200 words |
| 5 | My OpenClaw Configuration: A Complete Walkthrough | `idea` | openclaw.json deep-dive | ~2,500 words |
| 6 | How I Use Memory.md for Agent Continuity | `idea` | Memory system design | ~1,000 words |
| 7 | Telegram + OpenClaw: A ChatOps Love Story | `idea` | Channel integration guide | ~1,200 words |
| 8 | Cron Jobs That Actually Work: Lessons from the Trenches | `idea` | Heartbeat, maintenance patterns | ~1,000 words |
| 9 | The Family Assistant: Automating Calendar from Email | `idea` | Cloudflare Workers + CalDAV | ~1,500 words |
| 10 | When Agents Hallucinate Tools: Debugging Skill Failures | `idea` | Tool loop detection | ~1,000 words |

### AI News & Trends
*Analysis, opinions, and forward-looking thoughts*

| # | Idea | Status | Notes | Target Length |
|---|------|--------|-------|---------------|
| 1 | Local LLMs in 2026: What's Actually Practical? | `idea` | phi4, gemma4, qwen benchmarks | ~1,500 words |
| 2 | Why I'm Betting on "Sovereign AI" | `idea` | Local-first philosophy | ~1,200 words |
| 3 | The Multi-Agent Future: My 6-Month Experiment | `idea` | Socrates/Daedalus/Wadsworth retrospective | ~2,000 words |
| 4 | When Cloud LLMs Fail (And Local Saves the Day) | `idea` | Outage stories, fallback patterns | ~1,000 words |
| 5 | Building an AI Content Engine for Under $10/Month | `idea` | Cost optimization deep-dive | ~1,500 words |
| 6 | The Prompt Engineering Myth | `idea` | Opinions on "vibe coding" | ~800 words |
| 7 | Why I Don't Trust My Calendar to Google Anymore | `idea` | Privacy + automation trade-offs | ~1,000 words |
| 8 | Obsidian + AI: My Second Brain Setup | `idea` | Personal knowledge management | ~1,200 words |
| 9 | The Rise of "Agentic" Everything (And What's Real) | `idea` | Hype vs reality | ~1,000 words |
| 10 | What I Learned from Running 100% Local for a Week | `idea` | Challenge/experiment format | ~1,500 words |

---

## Content Pipeline Input

### Completed Milestones ✅
| Milestone | Status | Notes |
|-----------|--------|-------|
| **LocalAI Pipeline v1** | `completed` | Full content creation pipeline operational on Gaming PC (3080 Ti) |
| **Magic Wand UI** | `completed` | User input interface for content generation tested and working |
| **Dev Hardware Validation** | `completed` | End-to-end testing successful on local infrastructure |

### Next Up (Prioritized)
1. **Running Three AI Agents on a Beelink N150** — Socrates origin story, infrastructure deep-dive
2. **The Compaction Ghost Bug** — Technical post-mortem with lessons
3. **Building a Content Pipeline with Local LLMs** — Gaming PC 3080 Ti walkthrough (document the completed work)

### Future Goals — HoffDesk Expansion

#### Phase 1: Content & Pages (Q2 2026)
| Initiative | Status | Description |
|------------|--------|-------------|
| **Additional Site Pages** | `planned` | About, Projects, Uses (/uses), Now (/now) |
| **Content Archive** | `planned` | Searchable post index, category browsing |
| **RSS Enhancements** | `planned` | Category-specific feeds, JSON Feed |
| **Reading Time Estimates** | `planned` | Auto-calculated on all posts |

#### Phase 2: Design & UX Polish (Q2-Q3 2026)
| Initiative | Status | Description |
|------------|--------|-------------|
| **Design System Refinement** | `planned` | Unified typography, spacing, color tokens |
| **Dark Mode Perfection** | `planned` | System-aware toggle, persistent preference |
| **Mobile-First Optimization** | `planned` | Touch targets, responsive images, performance |
| **Accessibility Audit** | `planned` | WCAG 2.1 AA compliance verification |
| **Clean Overall Feel** | `planned` | Daedalus design audit implementation |

#### Phase 3: Media & Interactive (Q3-Q4 2026)
| Initiative | Status | Description |
|------------|--------|-------------|
| **Image Generation Pipeline** | `investigating` | Local FLUX/SDXL for article hero images |
| **Automated Thumbnails** | `planned` | Dynamic OG images for social sharing |
| **Code Syntax Highlighting** | `planned` | Theme-aware, copy-to-clipboard |
| **Interactive Diagrams** | `planned` | Mermaid, architecture visuals |

### Image Generation Investigation

**Status:** `researching` — See `/shared/project-docs/research/image-generation-pipeline.md`

**Options Evaluated:**
| Model | Location | Quality | Speed | Notes |
|-------|----------|---------|-------|-------|
| FLUX.1 [dev] | Gaming PC | High | ~30s/512px | Best quality, needs 12GB VRAM |
| SDXL Turbo | Gaming PC | Medium | ~2s/512px | Fast, good for thumbnails |
| Stable Diffusion 3 | Cloud API | High | Variable | Fallback for complex prompts |

**Decision:** Prioritize local FLUX.1 on 3080 Ti for hero images, SDXL Turbo for rapid iteration/thumbnails.

**Open Questions:**
- Prompt engineering for consistent style across articles?
- Storage strategy (S3 vs local vs Cloudflare R2)?
- Automated vs manual curation workflow?
- Attribution/metadata standards?

### Seasonal Themes
- **Spring 2026:** Infrastructure hardening, backup strategies
- **Summer 2026:** Family tech projects, automation wins
- **Fall 2026:** Year-in-review, tool evolution
- **Winter 2026:** Planning, architecture decisions for 2027

---

## Content Format Guidelines

| Element | Standard |
|---------|----------|
| **Tone** | Personal, conversational, "you're in my basement lab" |
| **Length** | 800–2,500 words (sweet spot: 1,200) |
| **Structure** | Hook → Problem → Journey → Solution → Reflection |
| **Code blocks** | Yes, with comments explaining "why" |
| **Images** | Screenshots encouraged, diagrams welcome |
| **Call-to-action** | "What's your setup?" or "How would you solve this?" |

---

## Pipeline Integration Notes

**For the Content Pipeline:**
- Read this file for topic selection
- Filter by `status: idea` for new content
- Respect `Target Length` for output sizing
- Check `Notes` for tone/angle guidance
- Update status to `in-pipeline` when work starts
- Update status to `published` with URL when live

**Tags for filtering:**
- `needs-research` — requires investigation before drafting
- `technical-heavy` — code, configs, architecture
- `personal-story` — narrative-driven, less tutorial
- `evergreen` — timeless, not news-dependent
- `timely` — tied to current events/releases

---

## Meta

**Document Maintenance:**
- Append new ideas to categories
- Update status as items progress
- Archive published posts with link
- Review monthly for stale ideas

**Last Pipeline Run:** —  
**Next Pipeline Run:** —  
**Active Pipeline Items:** —

---

*"The best posts come from the things that almost broke everything."* — HoffDesk Editorial Policy