TL;DR
I built a framework that measures two dimensions of AI collaboration: how sophisticated your tools are (Gas Town stages) and how deeply you're thinking (DOK levels). The intersection reveals where your next growth opportunity lives.
The Problem
Many AI practitioners face a hidden inefficiency: a mismatch between tool sophistication and task cognitive complexity.
| Anti-Pattern | Impact |
|---|---|
| Using powerful autonomous agents for simple "what is X?" queries | Unrealized potential |
| Asking deep strategic questions through basic chatbot interfaces | Bottlenecked thinking |
| No visibility into personal AI usage patterns | Stagnant growth |
| No framework for intentional growth in AI collaboration skills | Missed opportunities |
I noticed this pattern in my own work: I'd use powerful autonomous agents to answer questions I could have Googled. I was driving a Ferrari to the mailbox.
The tools were sophisticated. The questions weren't.
Without measurement, there's no improvement. I needed a mirror to see my AI collaboration patterns clearly.
Two Frameworks, One Insight
I combined two existing frameworks to create a two-dimensional model for understanding AI collaboration maturity:
Horizontal Axis: Gas Town Stages
Steve Yegge's "Welcome to Gas Town" (January 2026) describes an 8-stage progression of AI tool adoption:
| Stage | Name | Description |
|---|---|---|
| 8 | Full Gas Town | Complete AI-native development ecosystem |
| 7 | Agentic Workflows | Automated pipelines with agent coordination |
| 6 | Multi-Agent | Orchestrating multiple specialized agents |
| 5 | CLI Single Agent, YOLO | Terminal-based autonomous agent (e.g., Goose) |
| 4 | Chat IDE | Integrated chat in development environment |
| 3 | Copilot | Using AI code completion, inline suggestions |
| 2 | Curious | Experimenting with basic chatbots occasionally |
| 1 | Observer | Watching and evaluating AI tools, not yet actively using |
Most engineers I talk to are somewhere between Stage 3 and Stage 5. The tools are available. The question is: what are we doing with them?
Vertical Axis: Depth of Knowledge (DOK)
Norman Webb's Depth of Knowledge framework (1997) measures cognitive complexity across four levels:
| Level | Name | Prompt Indicators |
|---|---|---|
| 4 | Extended Thinking | "Research and synthesize...", "Create a framework...", "Investigate over time..." |
| 3 | Strategic Thinking | "Design...", "Analyze...", "What if...", "Develop a strategy..." |
| 2 | Application | "How would you...", "Compare...", "Explain why..." |
| 1 | Recall | "What is...", "List...", "Define..." |
DOK 1 is looking up syntax. DOK 4 is creating new knowledge. Most day-to-day engineering work lives at DOK 2.
The Integration Matrix
When you plot Gas Town stages against DOK levels, six distinct zones emerge:
1 DOK 1 DOK 2 DOK 3 DOK 4 2 (Recall) (Application) (Strategic) (Extended) 3 ┌──────────┬──────────────┬────────────┬────────────┐ 4Stage 6-8 │ Over- │ Over- │ Underutil- │ Frontier │ 5(Multi/ │ powered │ powered │ izing │ │ 6 Agentic) │ │ │ │ │ 7 ├──────────┼──────────────┼────────────┼────────────┤ 8Stage 5 │ Over- │ Underutil- │ Expected │ Growing │ 9(CLI YOLO) │ powered │ izing │ │ │ 10 ├──────────┼──────────────┼────────────┼────────────┤ 11Stage 3-4 │ Over- │ Expected │ Growing │ Frontier │ 12(Copilot/ │ powered │ │ │ │ 13 Chat IDE) │ │ │ │ │ 14 ├──────────┼──────────────┼────────────┼────────────┤ 15Stage 1-2 │ Expected │ Growing │ Thinking │ Thinking │ 16(Observer/ │ │ │ Ahead │ Ahead │ 17 Curious) │ │ │ │ │ 18 └──────────┴──────────────┴────────────┴────────────┘
The insight: Your tools and your thinking should grow together.
Zone Definitions
| Zone | Description | Action |
|---|---|---|
| Frontier | Pushing boundaries of both tool and cognition | Celebrate & Document |
| Thinking Ahead | High cognitive work with basic tools | Upgrade tools |
| Growing | Stretching into higher complexity, positive trajectory | Encourage |
| Expected | Appropriate match of tool sophistication to task complexity | Maintain |
| Underutilizing | Sophisticated tools for simpler tasks | Increase DOK |
| Overpowered | Tools exceed task needs—opportunity to level up your questions | Realign |
Every zone is valid. The framework isn't about judgment—it's about awareness and intentional growth.
What This Looks Like in Practice
I built a skill that analyzes AI collaboration sessions and returns a snapshot:
1╔══════════════════════════════════════════════════════════════════╗ 2║ SESSION ANALYSIS ║ 3╚══════════════════════════════════════════════════════════════════╝ 4 5GAS TOWN STAGE: 5 (CLI Single Agent, YOLO) 6 7DOK DISTRIBUTION 8──────────────────────────────────────────────────────────────────── 9DOK 1 (Recall): ████░░░░░░░░░░░░░░░░ 17% 10DOK 2 (Application): ████████████░░░░░░░░ 52% 11DOK 3 (Strategic): ██████░░░░░░░░░░░░░░ 26% 12DOK 4 (Extended): █░░░░░░░░░░░░░░░░░░░ 5% 13 14QUADRANT: Underutilizing 15──────────────────────────────────────────────────────────────────── 16You're using powerful autonomous tools—there's an opportunity to 17match your questions to that power. 18 19GROWTH NUDGES 20──────────────────────────────────────────────────────────────────── 211. Shift 2-3 DOK 2 prompts to DOK 3 by adding "analyze trade-offs" 222. Before simple queries, ask: "Can I make this more strategic?" 233. Try one DOK 4 extended investigation this week
The goal isn't to maximize DOK level on every interaction. Sometimes you need to look up syntax. The goal is awareness—knowing when there's an opportunity to go deeper.
📘 Want to try this yourself? Try the skill using
/rp-why initto establish your baseline,/rp-why currentto analyze your current session, or/rp-why compareto compare your session to your baseline.
The Growth Formula
Growth = (↑ DOK Level) + (↑ Gas Town Stage) + (Intentional Practice)
Moving from "Underutilizing" to "Growing" doesn't require new tools. It requires better questions:
| Instead of... | Try... |
|---|---|
| "How do I implement X?" | "What are the trade-offs between approaches A, B, and C for implementing X?" |
| "Fix this bug" | "Analyze why this bug occurred and what systemic changes would prevent similar issues" |
| "Write tests for this" | "Design a testing strategy that balances coverage, maintainability, and execution time" |
The tools are the same. The cognitive complexity is different.
Target User Profiles
The framework maps cleanly to different practitioner profiles:
| Profile | Typical Stage | DOK Distribution | Characteristics |
|---|---|---|---|
| Traditional | 1-2 | DOK 1: 60%, DOK 2: 30%, DOK 3: 10% | Minimal AI use |
| Adopter | 3-4 | DOK 1: 40%, DOK 2: 40%, DOK 3: 15%, DOK 4: 5% | Growing comfort |
| Practitioner | 5 | DOK 1: 25%, DOK 2: 45%, DOK 3: 25%, DOK 4: 5% | Autonomous agents |
| Advanced | 5-6 | DOK 1: 15%, DOK 2: 35%, DOK 3: 35%, DOK 4: 15% | Strategic use |
| Frontier | 7-8 | DOK 1: 10%, DOK 2: 25%, DOK 3: 40%, DOK 4: 25% | Agentic workflows |
The path from Practitioner to Advanced isn't just "use more tools." It's "do more DOK 3-4 work"—design systems, analyze trade-offs, create frameworks, drive technical direction.
AI tools can accelerate that path. Or they can let you coast at DOK 2 indefinitely. The framework helps you see which one is happening.
Actionable Takeaways
If you're in the Learning Zone (low stage, low DOK):
- This is a natural starting point—focus on learning the tools
- Try one new AI capability each session
- Don't worry about DOK yet—get comfortable first
If you're Overpowered (high stage, low DOK):
- Your tools exceed your task needs—opportunity to level up your questions
- Before each prompt, ask: "Can I make this more strategic?"
- Batch simple queries; save the agent for complex work
If you're Thinking Ahead (low stage, high DOK):
- Your thinking exceeds your tools—time to upgrade!
- Explore CLI agents or IDE integration
- Your DOK is strong; let better tools amplify it
If you're in the Frontier (high stage, high DOK):
- You're pushing boundaries—document what you learn
- Share patterns with others; teach what works
- Explore the edges: what's not yet possible?
Try It Yourself
Next time you're working with an AI tool, pause and ask:
- What Gas Town stage am I operating at? (Basic chat? Autonomous agent? Multi-agent?)
- What DOK level is this prompt? (Recall? Application? Strategic? Extended?)
- Is there a mismatch?
If you're using Stage 5 tools for DOK 1 questions, there's untapped potential. Small, consistent nudges compound into significant growth. The framework just helps you see where you are—and where you could go next.
Try the skill using
/rp-why initto establish your baseline,/rp-why currentto analyze your current session, or/rp-why compareto compare your session to your baseline.
Attribution
- Gas Town Stages: Steve Yegge, "Welcome to Gas Town" (January 2026)
- Depth of Knowledge: Norman Webb (1997). Criteria for alignment of expectations and assessments in mathematics and science education. Council of Chief State School Officers.
- Integration: The Gas Town × DOK integration is original to this work, combining these frameworks to create a two-dimensional model for reflective practice in AI collaboration.