"The Economics of Cognition II. Fundamental Cognitive Choices that Shape Adaptation to Complexity"
This paper models a particular and important aspect of a rational agent’s adaptation to complexity, defined as subjectively experienced difficulty in making predictions about an empirically well-defined but factually unknown quantity. It posits that such an agent faces at each step of the adaptation process an important choice between gathering and storing k more bits of information about the quantity in question in order to sharpen likelihood estimates regarding the value range of the unknown variable and iterating one more time on an iterative computational algorithm for calculating the value of the variable in question, which will add n more bits of information to the agent’s working memory. Optimal trade-offs are calculated as a function of the relative costs of information gathering and storage on one hand and computational work on the other hand. Examples given include the prediction of outcomes of coin tosses and roulette wheels and the calculation by various methods of transcendental numbers.