Every student who opens an AI and asks for help is making a decision they may not realise they are making: whether to think with the machine, or to let the machine think instead.
This book is about that decision.
Learning with AI, Thinking for Yourself introduces Mercurius - a pedagogical companion unlike any other. Mercurius never provides a direct answer. Instead, through a carefully structured series of dialogues, Mercurius returns every question to the student, names the moment when borrowed competence replaces genuine understanding, and holds the space in which real learning must occur.
The book is structured around the student's journey through five stages: Encounter, Dependency, Apprenticeship, Resistance, and Autonomy. Along the way, the reader meets archetypes drawn from contemporary university life - the business student who produces impressive presentations without understanding them, the data science student whose code works but whose understanding does not, the medical student who discovers that an AI can diagnose faster than they can think. Each encounter illuminates a different way in which artificial intelligence, used passively, can accelerate the production of work while quietly weakening the capacity for thought.
The book also contains the Gardner Sessions - fifteen conversations about mathematical paradoxes, from Hilbert's Hotel to the Prisoner's Dilemma, each chosen because it reveals a structural failure of human intuition. And the Logic Sessions, which address the hidden architecture of everyday argument: why "All S are P" cannot be reversed, how a valid argument can still be unsound, and where the error hides in a proof that appears airtight.
Throughout, the book is designed not only to be read, but to be loaded. Each chapter contains activation seeds - specific prompts that allow any AI system to adopt the Mercurius pedagogical method, transforming the book itself into a portable cognitive mentorship system.
Learning with AI, Thinking for Yourself does not argue against artificial intelligence. It demonstrates, conversation by conversation, what it means to remain intellectually present in its presence.
