AI & Causality

Towards Error Centric Intelligence I, Beyond Observational Learning

Modern AI systems are often evaluated by performance on observed tasks, but this can hide which errors are actually reachable, diagnosable, and correctable. This work asks whether progress toward general intelligence depends less on scale alone and more on the ability to expose, transform, and repair errors.

Marcus A. Thomas

This paper argues that progress toward artificial general intelligence is theory-limited rather than merely data- or scale-limited.

It develops an error-centric account of intelligence, emphasizing interventional competence, criticism, and the ability to transform unreachable errors into reachable and correctable ones.