•Mechanisms of phenotypic flexibility are ubiquitous in living organisms.•Some environmental variation is unpredictable, necessitating creative search procedures.•Iterative Bayesian updating is an attractive model of phenotypic flexibility.•Evolution can tune actual updating algorithms to approximate Bayesian updating.•The different mechanisms of phenotypic flexibility interact extensively with one another. Phenotypic flexibility includes systems such as individual learning, social learning, and the adaptive immune system. Since the evolution of genes by natural selection is a relatively slow process, mechanisms of phenotypic flexibility are evolved to adapt to contingencies on the time scales ranging from a few hundred milliseconds (e.g. avoidance of immediate physical threats) to a few millennia (e.g. cultural adaptations to local environmental variation in the Holocene). Because environmental variation is non-stationary and fat tailed, systems of phenotypic flexibility sometimes have to be creative. They do this by means of random innovation, or exploration, and selective retention. The canonically rational way to deal with variable, uncertain environments is the Bayesian process of using new data to update priors based on past experience. Organic evolution updates the gene frequencies of populations based upon the fitness of alleles. Learning updates behavioral priors based upon the reinforcement of alternate behaviors. Genes and mechanisms of phenotypic flexibility are not isolated but richly interact. Classically, genes are said to code for the reinforcers that shape behavior in individual learning, for example. It is currently controversial whether or not these interactions include a role for the products phenotypic flexibility directly shaping selection on genes.