Skip to content

growth_pipeline_plan.md

Latest commit

 

History

History
115 lines (91 loc) · 5.32 KB

growth_pipeline_plan.md

File metadata and controls

115 lines (91 loc) · 5.32 KB

Personal Growth Pipeline Plan — AI Dev Account

Overview

A parallel pipeline to create longer, meaningful X posts (threads, case studies) without interfering with the existing news-digest pipeline.

1. Growth Foundations for ~10-Follower AI-Dev Account

What Works Why It Moves the Needle How to Codify
Consistent visual branding Feed looks cohesive, higher follow rate Pick static banner + 2-color palette
Signature rhythm Algorithm rewards regular activity Post at same 3 times daily (when ready)
Micro-CTA at end of threads Turns readers into participants Auto-append 2-sentence CTA block
Thread length 3-6 tweets Optimal completion rate Stop at 5 tweets unless manually extended
Reply-to-self updates Keeps conversation open Schedule "result tweet" 24h later

2. High-Level Architecture (Parallel Pipeline)

┌─────────────────┐      ┌───────────────────────┐
│  Source Store   │      │   Experiment Vault    │
│  (your notes)   │      │  experiments_inbox/   │
└───────┬─────────┘      └───────┬───────────────┘
        │                          │
        ▼                          ▼
   ┌─────────────────┐    ┌───────────────────────┐
   │   Content-Prep │    │   Scoring & Routing  │
   └───────┬─────────┘    └───────┬───────────────┘
           │                      ▼
           ▼                ┌───────────────────────┐
      ┌───────────────┐       │   Thread Builder      │
      │   Generator   │       │   (4-5 tweet skeleton)│
      └───────┬───────┘       └───────┬───────────────┘
              │                       │
          ┌───────┴───────┐       ┌───────┴───────┐
          │   Validation  │       │   CTA Injection│
          └───────┬───────┘       └───────┬───────┘
                  │                       ▼
                  ▼                ┌───────────────────────┐
          ┌───────────────┐        │   Telegram Sender   │
          │   Dry-Run     │        │   (dry-run first)  │
          │   (txt dump)  │        └───────┬───────────────┘
          └───────┬───────┘                │
                  │                        ▼
                  ▼               ┌───────────────────────┐
          ┌─────────────────┐      │   Post-Log Append   │
          │   Manual Review │      │   (posted_log.txt)   │
          └─────────────────┘      └───────────────────────┘

3. Experiment Vault Structure

/personal_experiments/
   2026-10-25_exp_grok-suggests-thread/
   ├─ meta.txt          # title, date, status, tags
   ├─ notes.txt         # step-by-step description
   ├─ screenshots/      # optional proof
   ├─ results/          # benchmark numbers
   └─ voice.md          # tone guidelines

4. Thread Generation Workflow

  1. Folder Scan – list newest meta.txt folders
  2. Scoring – compute first-party strength (high/medium/low)
  3. Lane-Keyword Filter – must match ≥2 keywords
  4. Visual Confirmation – check visual requirements
  5. Thread Builder Prompt – LLM generates 4-5 tweet thread
  6. CTA Injection – append call-to-action
  7. Dry-Run Save – write to personal_posts/dry/
  8. Manual Review – quick sanity check
  9. Publish – send to X + schedule follow-up
  10. Metrics Capture – track impressions/engagement

5. Separation From Existing Pipeline

Aspect Existing Pipeline New Pipeline
Source RSS feeds, public news Your experiment vault
Content Type 1-2 sentence snippets 4-5 tweet threads
Scoring Freshness & dump strength First-party strength
Output Shared archive/ logs Own personal_posts/ folder
Automation Fully autonomous Semi-autonomous (manual check)

6. Next Steps

  1. Map your Grok recommendations to this framework
  2. Tailor lane-keywords to your specific topics
  3. Set up skeleton scripts (scan, score, builder)
  4. Create first experiment folder template

7. Your Input Needed

  • Visual style preferences?
  • Topics/keywords for "in-lane"?
  • Tone constraints beyond dry humor?
  • Preferred posting cadence?

Ready to implement once you provide your Grok suggestions and preferences!