AI Safety Researcher Compiler

A comprehensive, interactive curriculum for systematically developing AI safety research capabilities - from foundations to advanced technical contributions

Two Modes, One Journey

🎓 Academic Mode

Rigorous, technical content with formal definitions, mathematical proofs, and citations. Presenting established research, consensus positions, and empirical findings from the AI safety literature. This mode maintains academic neutrality - focusing on what we know, how we know it, and where uncertainty remains.

💭 Personal Mode

My personal takes on AI's most disputed questions. Covering impact, goals, and responses where evidence allows multiple views. These opinionated perspectives invite challenge and revision - modeling the intellectual humility and value pluralism this project advocates.

Toggle between modes anytime to match your learning style and goals

Value Pluralism

Acknowledging multiple valid approaches to AI safety

Hands-on Implementation

Build transformers and alignment techniques from scratch

Progressive Competency

Embedded assessments verify understanding at each stage