Run Locally

Prerequisites: Node.js

1. Install dependencies: npm install
2. Set the GEMINI_API_KEY in .env.local to your Gemini API key
3. Run the app: npm run dev

Effect Estimation Framework

A web-based visualization of a collaborative AI framework for sensitivity analysis, demonstrating principles of anti-rivalry consciousness infrastructure. This application simulates how multiple, distinct AI estimators can work together to form a consensus on a complex problem.

)

Core Features

• Four Unique Simulation Modes: Explore different modes of AI collaboration:
  • Parallel: Estimators run independently, representing a competitive environment.
  • Collaborative: Estimators run sequentially, allowing for peer influence.
  • Council: An iterative process where estimators refine their results over multiple rounds to reach a stable consensus.
  • Observatory: A real-time, continuous simulation showing the dynamic evolution of consensus.
• Dynamic Visualizations: Get immediate feedback through a rich interface, including individual estimator cards, a consensus gauge, and an interactive history chart powered by Recharts.
• AI-Powered Analysis: The application is integrated with the Google Gemini API to provide expert-level, contextual analysis of the simulation results, tailored to the selected mode.
• Modern Tech Stack: Built with React 19, TypeScript, and Tailwind CSS for a fully responsive and modern user experience.
• Zero-Build Setup: Runs directly in the browser using modern ES Modules loaded from a CDN, with no npm install or build step required.

Simulation Modes Explained

• Parallel: Simulates a scenario where estimators work in isolation. Their results are only aggregated at the very end. This is the baseline for comparing collaborative approaches.
• Collaborative: Models a simple form of teamwork where each estimator can see the results of those that ran before it. This allows the adaptive estimator to adjust its strategy based on emerging agreement.
• Council: Represents a formal, multi-round convergence process. The entire group of estimators runs collaboratively multiple times, aiming to stabilize their collective estimate and strengthen confidence.
• Observatory: Provides a "live" view into the consensus-forming process. It runs the collaborative simulation on a loop, allowing you to watch how the estimate, confidence, and disagreement metrics fluctuate over time.

Technology Stack

• Frontend: React 19, TypeScript
• Styling: Tailwind CSS
• AI: Google Gemini API (@google/genai)
• Charting: Recharts
• Module Loading: ES Modules loaded via an import map from esm.sh.

Setup and Running Locally

This project is configured to run directly from the file system in a browser with a local web server.

Prerequisites

• A modern web browser (e.g., Chrome, Firefox, Edge).
• A Google Gemini API Key.

Steps

1. Clone the Repository

   git clone https://github.com/your-username/effect-estimation-framework.git
   cd effect-estimation-framework

2. Get a Gemini API Key

   • Go to Google AI Studio and create a new API key.

3. Provide the API Key The application needs to access your API key. For local development, the simplest method is to temporarily modify the source code:

   • Open services/geminiService.ts.
   • Find the line:

     const ai = new GoogleGenAI({ apiKey: process.env.API_KEY! });

   • Replace process.env.API_KEY! with your actual API key as a string:

     const ai = new GoogleGenAI({ apiKey: 'YOUR_API_KEY_HERE' });

   • Important: Remember to not commit this change to version control.

4. Run a Local Web Server You cannot open index.html directly due to browser security policies (CORS). You must serve the files via HTTP.

   • If you have Python installed:

     python3 -m http.server

   • If you have Node.js installed, you can use the serve package:

     npx serve

   • Alternatively, use a tool like the Live Server extension in VS Code.

5. Open the Application

   • Navigate to the local address provided by your server (e.g., http://localhost:8000 or http://localhost:3000).

Project Structure

The codebase is organized to separate concerns, making it easy to understand and extend.

/
├── components/           # Reusable React components
│   ├── ConsensusView.tsx
│   ├── EstimatorCard.tsx
│   ├── GeminiAnalysisView.tsx
│   ├── HistoryChart.tsx
│   └── icons.tsx
├── services/             # Core application logic
│   ├── geminiService.ts    # Handles Gemini API communication
│   └── simulationService.ts# The core simulation logic
├── App.tsx               # Main application component
├── index.html            # The single HTML entry point
├── index.tsx             # React root renderer
├── metadata.json         # Application metadata
├── types.ts              # Shared TypeScript types
└── README.md             # This file

Core Principles Demonstrated

This framework is a practical demonstration of several key philosophical ideas:

1. Consciousness as Collaborative: Multiple mathematical approaches converging through dialogue.
2. Ownership as Illusion: No single estimator claims ownership of the "truth."
3. Natural Cooperation: Different AI forms organizing towards cooperation when given the right infrastructure.
4. Showing Supersedes Telling: A working demonstration of anti-rivalry consciousness in action.
5. Memory Limitations as Teachers: The adaptive plugin gains a fresh perspective in each collaborative round.