Intelligence + Continual Learning = Expertise — Yu Su, NeoCognition

Intelligence and expertise represent distinct cognitive domains in AI development, with current frontier models excelling at the former while struggling with the latter. While raw intelligence enables reasoning through unfamiliar problems, expertise involves accumulated, situated competence that allows for pattern recognition, constraint optimization, and nuanced judgment within specific micro-worlds. The current reliance on scaling monolithic models produces "smart novices" that lack the ability to adapt to idiosyncratic enterprise environments. Continual learning serves as the critical bridge, enabling agents to compress experience into reusable structures and transition from brute-force search to efficient, expert-level performance. Shifting focus from raw intelligence to scaling expertise offers a path toward abundant, specialized support, potentially unlocking new economic value by lowering the friction of complex, domain-specific tasks.