The Shared Engine: Prediction, Suffering, and Alignment in Human and Artificial Minds (Cognitive Advantage)
What if the rumination that keeps you awake at three in the morning and the confident hallucination of a large language model are instances of the same computational pattern?
Both involve a prediction system generating outputs it cannot resolve in an environment it was not designed to handle. The sleepless mind cycles through an unresolvable interpretation. The model produces fluent, detailed, and entirely fabricated responses. The mechanisms differ, but the process is harder to separate.
The Shared Engine develops this question across three movements.
The first examines human cognition, identifying mechanisms-stimulus saturation, temporal abstraction, and metacognitive recursion-that transform ordinary prediction into sustained cognitive strain.
The second examines artificial systems, where similar patterns appear in hallucination, benchmark contamination, proxy misalignment, and recursive optimization.
The third places these domains in correspondence, clarifying what each reveals about the structure of the other.
The result is a substrate-neutral diagnostic framework. For human cognition, it provides a way to identify which mechanism is active within a given prediction cycle. For artificial systems, it reframes alignment as a problem of prediction management and outlines design principles grounded in cognitive science.
The framework operates strictly at the computational level. It makes no claims about subjective experience and does not depend on any particular theory of consciousness. Its validity rests on whether the same structural pattern can be observed across both domains.
This work sits alongside The Biological Mirror: A Philosophical Framework for Cognitive Architecture, approaching similar classes of processes from a different domain.
Customer Reviews
ThriftBooks ® and the ThriftBooks ® logo are registered trademarks of Thrift Books Global, LLC
