Comparison v3: AI Tools in the Agentic Era

Interactive comparison tool for AI services using Ethan Mollick's Models -> Apps -> Harnesses lens.

Built with vanilla HTML/CSS/JS, no build step, and designed for GitHub Pages deployment.

Quick Start

1. Install dependencies:

   pnpm install

2. Install Playwright browsers if this is a fresh machine:

   pnpm exec playwright install

3. Run tests:

   pnpm test

4. Optional local preview:

   python3 -m http.server 8000

   Open http://localhost:8000.

Current App Structure

• index.html -> v3 UI and all styles
• app-v3.js -> data model, filtering/sorting/toggles, rendering logic
• tests/v3.spec.js -> Playwright E2E + accessibility + mobile tests
• v1/ -> archived 2025 version

Key Features

• Category pill filtering (Chatbot, Coding Agent, Work Agent, Research Tool)
• Boolean-row click sorting (true-first / false-first)
• Service and attribute visibility toggles
• Fuzzy search with Fuse.js
• Harness strength dot visualization
• Theme toggle with persisted light/dark mode
• Mobile form-factor UX:
  • touch-friendly control sizing
  • swipe hint for table comparison
  • focus one service mode (next/prev + selector)

Scripts

• pnpm test -> run Playwright suite (Chromium)
• pnpm test:headed -> run tests with visible browser
• pnpm test:report -> open last Playwright report

Notes

• Playwright uses port 8002 via playwright.config.js.
• Keep PII out of committed content.
• Main data source: Mollick Feb 2026 guide + project-specific updates.