2025-11-06_Mastermind

2025-11-06 AI Mastermind for Leaders

Table of Contents

• Insight - Your Knowledge Is the Database, AI Is the Interface
• Insight - Persistence Through the Dip Is the Real Competitive Advantage

Session Overview

The November 6 session opened with two significant member wins that set the tone for everything that followed. Dirk announced he had successfully set up his own self-hosted N8N server in Europe — a milestone several months in the making, achieved through persistence through repeated technical obstacles. Lou’s response — “The 4-minute mile has been broken, so let’s do it!” — was both celebration and invitation. Don Back shared that he used Claude’s deep research function to complete 5 days of grant research in a fraction of the time, producing output that left his university client “almost incredulous.”

These wins were not incidental — they were the lived demonstration of the mastermind’s core thesis. The people who push through the dip arrive at capabilities that look remarkable to those who gave up earlier. The session used both examples as launch points for a deeper conversation about knowledge infrastructure: what happens when you stop treating your expertise as something in your head and start treating it as something in a database.

Lou then walked the group through his Qdrant/OpenWebUI/ChatGPT integration — the system he had built for the Jansen legal team — as a template for what any coach or knowledge entrepreneur could do with their own accumulated content. The demonstration connected Don’s Notebook LM instinct with the more scalable architecture of a vector database connected to a conversational AI interface.

High-Signal Moments

• Dirk’s server achievement generated one of the most energizing moments of the series: a concrete, personal proof point that the dip is survivable and what’s on the other side is genuinely different
• Don’s grant research story was a precise measurement of AI impact: task that normally takes a week done in a fraction of the time, with clients surprised by the turnaround
• Don articulated the knowledge-base insight spontaneously: “What I’m actually doing is pulling together whatever my brain is in this area” — recognizing that the Notebook LM plan was not just practical but a first step toward externalizing his expertise
• Lou’s Qdrant demo showed the full stack: documents uploaded to a vector database → MCP or ChatGPT action created to access it → natural language queries returning semantically relevant results from the full document set
• The local vs. cloud database choice was explicitly named as a privacy consideration — coaches and consultants with sensitive client data can run a full vector database locally without any cloud exposure
• Kasimir reminded the group to always use the latest prompt versions (his prompts carry version numbers for exactly this reason) — a small but important systems hygiene point
• Lou used the metaphor of “automation of engagement” — the AI is automating the brainstorming and writing process, not replacing the thinking

Open Questions

• What is the right size and format for a knowledge base to be genuinely useful — does it need to be comprehensive, or does curation matter more?
• How do you handle the privacy and consent considerations when putting client-related materials (even anonymized) into a retrieval system?
• What is the minimum viable knowledge base for a coach who wants to build a client-facing AI tool based on their methodology?
• How does semantic retrieval compare to a well-organized Notion or document system — what is the actual productivity difference?
• When does a personal knowledge base become a productizable asset — what has to be true about its quality and coverage?

Suggested Follow-Through

• Identify your 5 most important knowledge assets (transcripts, frameworks, materials) and put them into a Notebook LM or Perplexity Space as a first step toward a retrievable knowledge base
• For those with privacy-sensitive materials: explore the Qdrant local installation — run one test with a small document set before committing to the full migration
• Don: proceed with the knowledge base plan for the University of Alberta program design; the depth and specificity of the retrieval will make the program distinctly yours
• Dirk: celebrate the server win and immediately move to the next N8N flow — Supabase integration is the logical next step
• Document one instance this week where you push through a dip instead of stopping — note what it felt like before and after

Additional Resources

Links & Tools Shared in Chat

• Cozora — shared by Ri Ca (copied from Lou’s Telegram message); context not specified in chat

Books & Articles Mentioned

• None.

Ideas from Chat

• Rich Schefren’s AI Mastermind and Jay AI: Ri Ca mentioned being in Rich Schefren’s AI mastermind (monthly meetings with Rich directly plus an AI expert). Rich Schefren’s AI expert built “Jay AI.” Ri Ca noted the “Rich Schefren Brain” feature is an extra monthly fee. Worth monitoring as a comparable high-level AI mastermind model.
• Claude as a side consultant: Don Back — “I can get started (with Claude open on the side for consults)” — a practical habit pattern worth naming: using an AI model as a real-time technical consultant during implementation sessions.