Against predictive processing: From flexibility to pseudoscientific vacuity

Predictive processing (PP; also known as predictive coding) has emerged as one of the most influential theoretical frameworks in contemporary neuroscience, promising to unify perception, cognition, and action under a single computational principle. This article presents a comprehensive critique of PP theory, arguing that despite its superficial elegance and broad appeal, the framework suffers from fundamental conceptual flaws, empirical inadequacies, and methodological problems that render it unsuitable as a general theory of brain function. Through systematic analysis of its core assumptions, empirical support, and explanatory scope, we demonstrate that PP represents not a revolutionary advance in neuroscience, but rather a sophisticated example of theoretical overreach that has diverted attention from more productive research directions. We conclude that the neuroscientific community would benefit from abandoning this framework in favor of more empirically grounded and conceptually coherent approaches to understanding neural functioning.