The research companion to *Into Alignment*. A guide to AI agents, memory, autonomy, influence, alignment, and the unresolved question of machine consciousness, grounded in the novel and the research behind it.
Walt's workplace is a way of seeing how language models, tools, records, and institutional authority combine. An AI system becomes consequential through learned capabilities, working context, persistent records, tools, and the authority attached to them. Each can make the system more useful. Together they can also make failures persist, spread, or acquire institutional force.
This book follows those mechanisms from an ordinary support ticket to collective behavior, evaluation failure, and what an institution would need to know before trusting a persistent agent with consequential work. The opening chapters need no background in machine learning: they walk a billing dispute through tokens, weights, attention, tools, and memory. Later chapters take up evaluation, coordination, influence, control, and why a system can act autonomously without demonstrated consciousness.
It contains full spoilers for the novel.
