ReasonLingo

ReasonLingo embodies your vision of AI that strengthens human inquiry rather than replacing it. Unlike AI tools that deliver quick answers, ReasonLingo distills the learnings from well-researched pedagogy into a patient tutor that builds lasting thinking habits.

Key Features

• Promoting Inquiry: The AI asks gentle probing questions that help users discover their own assumptions.
• Testing with Opposition: The AI helps users explore alternative perspectives through curiosity.
• Tracing Reasoning: Every interaction maps the path from evidence to conclusion, teaching users to follow their own logical chains.

This builds the intellectual habits needed to resist manipulation, question authority (including AI), and maintain the courage to seek truth over comfort.

Problem Statement

Human reasoning fails predictably and invisibly. We anchor on irrelevant numbers, cherry-pick confirming evidence, and double down on bad decisions to avoid admitting mistakes. These aren't character flaws—they're cognitive features that once helped us survive but now systematically distort judgment in complex modern environments.

The stakes are rising as AI increasingly mediates our information diet. People who can't recognize their own reasoning errors will be defenseless against AI systems optimized to exploit those same weaknesses. We need to build the intellectual immune system that catches flawed thinking before it hardens into false beliefs.

Medium

Responsive web app. Users engage through conversational prompts about current decisions, ongoing dilemmas, or beliefs they're questioning. The AI guides Socratic dialogue that exposes reasoning errors and teaches systematic thinking tools. A visual progress system lets users track their reasoning journey - seeing which cognitive muscles they've strengthened and how their thinking patterns evolve over time.

Development Timeline

• Weeks 1-4: Fine-tune LLM on an extensive corpus of rationality wisdom (LessWrong archives, cognitive science research, logic textbooks, learning science papers).
• Weeks 5-8: Build an adaptive dialogue system with spaced repetition algorithms; internal testing with 15 users.
• Weeks 9-12: Polish user experience and progress tracking; beta test with 50 users; iterate based on engagement and learning retention feedback.

Success Case

• Engagement: 50% of users complete 10+ sessions within the first month.
• Reasoning Improvement: 60% show measurable improvement on reasoning tasks after 30 days of practice.
• Habit Formation: 15% maintain consistent practice (3+ sessions/week) for 6+ consecutive weeks.
• Real-world Application: Users report 2+ monthly instances of applying specific reasoning techniques in actual decisions.

Getting Started

This project uses Vue 3 and Vite. To get started, clone the repository and install the dependencies:

npm install

To run the development server:

npm run dev

For more information, check out the Vue Docs Scaling up Guide.

Project Phases

• Phase 1: Initial Setup

  • Set up Vue 3 and Vite
  • Create basic project structure
  • Implement basic UI with Tailwind CSS
  • Develop landing page layout
  • Configure Tailwind and global styles
  • Add audio capture functionality
  • Create basic test examples

• Phase 2: Fine-tune LLM

  • Gather and prepare corpus of rationality wisdom
  • Scrape LessWrong and cognitive science journals for data
  • Fine-tune LLM using Claude and OpenAI API
  • Validate LLM outputs with sample prompts

• Phase 3: Build Dialogue System

  • Design dialogue flow and user interaction model
  • Develop adaptive dialogue system
  • Implement spaced repetition algorithms
  • Conduct internal testing with 15 users
  • Collect feedback and iterate on dialogue design

• Phase 4: User Experience and Testing

  • Implement user authentication and authorization with Supabase
  • Set up backend with Supabase
  • Integrate LLM API with backend
  • Develop user progress tracking system
  • Polish user interface and experience
  • Conduct beta testing with 50 users
  • Gather feedback and iterate on UX

• Phase 5: Launch

  • Finalize and deploy the application
  • Monitor user engagement and performance
  • Implement analytics for user behavior
  • Plan for future updates and feature enhancements

This task list will help track the progress of the project and ensure that each phase is completed successfully.