This book develops the paradigm that any complex entity must have an accurate internal model to inform it of the available resources and opportunities given its nature and environment. Such knowledge is necessary to make the entity's complexity compatible with the second law. The model's accuracy may only be maintained through Bayesian inference; updating the model upon the reception of new data or experience in accordance with Bayes' theorem. I argue further that Darwinian processes are a physical implementation of Bayesian inference. In quantum systems the internal model takes the form of...