Provides François Chollet's framework for understanding intelligence, AGI development paths, and the limitations of current AI approaches. Use this skill when users ask about- (1) What intelligence really means and how to define AGI, (2) Why scaling pre-training alone won't achieve AGI, (3) The difference between memorized skills and fluid intelligence, (4) Test-time adaptation and its role in AGI, (5) The ARC benchmark and what it measures, (6) Type 1 vs Type 2 abstraction in AI systems, (7) Program synthesis approaches to intelligence, (8) Evaluating claims about AGI progress, or (9) Understanding the conceptual foundations needed for building generally intelligent systems.