Part 13/16:

He observes that science currently lacks a solid framework for comparing diverse tasks in terms of their relation to intelligence. Without a theory of tasks, measuring and guiding progress remains challenging. Moving forward, Thoren advocates prioritizing novelty as a proxy for generality—an AI’s ability to handle unseen, unpredictable situations—and forming systems capable of reflecting on their own learning process.

Scientific Cycles and the Journey of Innovation

Thoren describes scientific advancement as cyclical, involving periods of intense effort, reflection, and sudden insights—epiphanies—that propel the field forward. This process underpins how foundational ideas emerge and mature, eventually transitioning from research labs to real-world applications.