Skip to content

increasinglyHuman/ai-role-evolution

Repository files navigation

CONDUCTOR: The Collaboration Gradient

Play Now License: MIT Built With

"Eighteen months ago I was afraid the AI would delete my database. Yesterday I realized I'm now the more likely source of danger to the application."


What This Is

CONDUCTOR is an interactive self-discovery experience that helps professionals find where they actually are on the human-AI collaboration gradient — and more importantly, see clearly what the next step looks like for them, in their role, in their real work.

It is not a quiz. It is not a personality typing system. It is not a marketing tool.

It is a mirror. Built with enough precision that looking into it tells you something true.

Built as a companion piece to the LinkedIn article: "The Shift Nobody's Teaching: From Doing With AI to Thinking With AI" by Allen Partridge, Director of Digital Learning Product Evangelism, Adobe.


The Five Stages

Stage Identity The Work
Passenger "Do this for me." Transactional. Request → output → edit.
Pilot "Let me steer this." Iterative. Prompt craft. Reliable techniques.
Co-Author "Let's build this together." Collaborative. Work emerges from the relationship.
Architect "I define the why. You handle the how." Intent specification. Human experience first.
Conductor "I hold the vision. The ensemble plays." Multi-workstream orchestration. Vision as the entire value.

These stages are not a ladder of competence. They describe where your value lives in the human-AI workflow. Most people move fluidly across multiple stages depending on project scale and context. The skill chain applies to a quick afternoon solution just as much as a massive long-form project.


The Experience

Role Selection — Three tracks: L&D / Instructional Designer, Developer / Technical, Manager / Leader / Strategist. Role selection genuinely changes scenario language throughout — not just the output page.

7 Scenario-Based Situations — Not "how good are you at X" but real work moments. Multi-select: because in real professional life you carry multiple competing instincts simultaneously. The pattern of what you select together is more diagnostic than any single choice.

Attention Capital — A finite resource that makes choices feel consequential. Where you instinctively don't want to spend attention is itself diagnostic.

Living Gradient Visualization — A persistent visual showing your personal distribution across the five stages, updating in real time as you move through the experience.

Micro-Reveals — After each scenario, a brief contextual reflection surfaces. Not judgment. Illumination. You draw your own conclusion.

Mission Debrief — Your stage profile, rendered as a distribution. What it looks like from where you are. The one practice worth starting tomorrow. What intentional infrastructure looks like for this kind of development.


Architecture

src/
├── main.js                   ← Entry point
├── style.css                 ← Mobile-first styles
├── state.js                  ← Reactive state + gradient distribution vector
├── data/
│   └── scenarios.js          ← All content: 3 tracks × 7 scenarios + debrief
├── ui/
│   ├── background.js         ← Animated background
│   ├── scenes.js             ← Scene visuals per scenario
│   └── screens.js            ← Screen transitions
└── game/
    ├── scenario.js           ← Scenario rendering + multi-select mechanic
    ├── attention.js          ← Attention Capital budget system
    ├── gradient.js           ← Live gradient visualization + distribution scoring
    ├── reveal.js             ← Micro-reveal logic after each scenario
    └── debrief.js            ← Mission debrief + output generation

Zero dependencies. Vanilla JS + Vite. Mobile-first. Target: under 40KB gzipped.


Key Design Decisions

Multi-select scoring uses weighted pattern recognition across the full response set — not arithmetic. The output is a distribution range, not a number. "Operating primarily between Co-Author and Architect, with strong Pilot instincts still active" is more accurate and more useful than a single label.

No single correct answer — The experience respects the reader's intelligence. There is no trick. Selecting Conductor-stage answers doesn't make you win. Authentic responses produce useful output. Gaming it produces noise.

The Attention Capital mechanic is a reflective signal, not a punishment system. Low capital doesn't block progress — it surfaces something true about where you're spending yourself.

Role routing is handled entirely in the data layer. No conditional branching inside render logic. scenarios.js contains all three track variants. state.js gates content based on selected role throughout.


Scoring Model

Each answer option carries a stage weight vector across all five stages. After each scenario, the player's selection pattern updates a running distribution in state.js. At debrief, gradient.js renders this distribution as a visual spectrum — the player's personal mass shown across the gradient.

Example weight vector (Architect-primary answer with Co-Author resonance):

weights: { passenger: 0, pilot: 0.1, coauthor: 0.3, architect: 0.8, conductor: 0.2 }

The distribution visualization is the centerpiece of the final debrief — specific enough to screenshot and share.


Intellectual Roots

  • Allen Partridge — lived experience of the Conductor arc; the ADR/UX gap insight; the philosophy that the best human-AI collaboration happens when both parties can surprise each other
  • Nate B. JonesPrompting Just Split Into 4 Different Skills (Feb 2026) — complementary framework naming the prompting discipline taxonomy; CONDUCTOR names the identity transformation underneath
  • EY Work Reimagined 2025 — 88% AI adoption, 28% transformation readiness
  • McKinsey State of AI 2025 — the redistribution pattern
  • IBM Institute for Business Value 2025 — 1 in 4 AI projects delivers promised ROI
  • Josh Bersin Company 2026 — 74% of companies not keeping up with their own demand for new skills
  • PwC Global AI Jobs Barometer 2025 — 25% wage premium for AI skills, rising
  • Klarna 2024-2025 — the most instructive case study in the field; the rehire arc as evidence that value always lived, and remains, in the human judgment layer

Development

npm install
npm run dev
npm run build
npm run preview

Credits

Concept & Direction: Allen Partridge (@doctorpartridge) Development: Claude (Anthropic) — architecture, game logic, scoring system, CSS, module structure Scene Art: Generated with Meshy AI Icons: Brutalist Bob icon set Inspired by: PATHFINDER: First Contact Protocol — poqpoq.com/adobe/pathfinder/


Built for the LinkedIn learning community. Designed to make you see yourself clearly — in the best way.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors