AI as Your Partner
Project Overview
Status: Active
This project explores two related questions about working with AI tools:
- How do you configure AI to work well? (mechanics)
- What does it mean to partner with AI? (philosophy)
The first question led to a published article on rules files. The second emerged from trying to understand why those rules work.
Two Threads
The Mechanics: What Makes Good Rules Files
How to structure AI instructions, where to put them, what to include.
- Rules Research Summary — The original research on universal vs. personal rules, the spectrum model, format convergence
- What Makes Good AI Rules Files — The published article
Key insight: Format converges (everyone needs the same sections), but content remains personal (how you fill them out reflects your values and workflow).
The Philosophy: What is Partnership
What makes collaboration with AI feel like partnership rather than delegation.
- What is Partnership? — Working notes on the philosophy underneath the rules
- Research: AI Partnership (Claude) — Academic/philosophical synthesis
- Research: AI Partnership (ChatGPT) — Practitioner-focused synthesis
- Research: Human Partnership Patterns — Cross-domain patterns from marriage, business, creative, and high-stakes partnerships
Key insight: With AI, you’re not just defining the partnership contract. You’re defining the partner. Rules files don’t negotiate with an existing entity; they shape what entity shows up.
The Connection
The mechanics article answers “how do I write good rules?” The philosophy work answers “why do those rules create something that feels like partnership?”
The answer: partnership emerges when you teach the AI how to think about work, not just what work to do. “Research before acting” isn’t a task boundary; it’s a reasoning pattern. “Never skip tests” isn’t scope; it’s a value. The rules are training, not just configuration.