Mar 18 • Weekly Session Notes
AI Automation Deep Dive - Feedback Loops, Crons & Tech Stack
Slide deck from the AI Mastermind group covering the full automation system architecture.
Key topics:
1. The Setup - Two AI agents working together: CASE (Mac Mini) handles content, outreach, and infrastructure. TARS (MacBook) handles operations, client work, and pipeline. Both share data via git sync.
2. Feedback Loop Architecture - The core concept: Data Collect > Analyze Patterns > Rewrite Strategy > Execute > repeat. Active loops include: Content Performance, LinkedIn Outreach, Engagement Rules, Blog Analytics, Profile Viewers, Calendly Attribution.
3. Daily Schedule - Morning: LinkedIn acceptance check, content ideas. Afternoon: DM replies, engagement. Evening: Content scheduling, pipeline review. Overnight: Batch crons, git sync.
4. Model Allocation - Opus for strategy/conversation (main session only), Sonnet for human-facing tasks, Qwen/MiniMax for mechanical/template tasks. 46 of 57 crons run on free/cheap models.
5. Tech Stack - OpenClaw (agent framework), Claude, MiniMax, Qwen 3.5, Facebook Graph API, Instagram API, Calendly API, Himalaya email CLI, git-based sync between machines.
6. Key Learnings - Sub-agent rules, Ralph loops for long-running tasks, browser profile isolation, content performance tracking, autonomy guardrails.
Slide deck available - reply if you want the link.
0
0 comments
AI Automation Deep Dive - Feedback Loops, Crons & Tech Stack
powered by
skool.com/ai-mastermind-3877
Weekly AI mastermind for builders, founders & operators. Real conversations about using AI to grow your business.
Suggested communities
Powered by