Pathological decision-making
Palminteri and Wyart's analysis of pathological decision-making reveals the computational foundations beneath everyday judgment by examining what happens when these systems fail.
Their focus on Bayesian inference breakdown in psychosis and the shift from goal-directed to habitual behavior in addiction provides a natural laboratory for understanding normal decision-making processes.
The work connects directly to Herbert Simon's bounded rationality — showing not just that our reasoning is limited, but how specific constraints shape different failure modes.
For product leaders, this offers insight into why users make seemingly irrational choices and how digital environments might inadvertently exploit the same vulnerabilities that create pathological patterns.
The clinical perspective grounds abstract decision theory in observable human behavior.
Central argument
Palminteri and Wyart argue that pathological decision-making — as seen in psychosis and addiction — is not merely dysfunctional behavior but a window into the computational architecture that governs normal judgment. Their central thesis is that Bayesian inference breaks down in psychosis in identifiable ways, while addiction represents a measurable drift from goal-directed to habitual control systems. By mapping failure modes, they reverse-engineer the underlying mechanisms of healthy cognition, positioning clinical pathology as an empirical method for testing decision theory rather than an outlier to be explained away.
Critique
The framework's explanatory power depends heavily on the assumption that pathological states represent clean degradations of normal computational processes — but this may understate how qualitatively different clinical conditions are from everyday bounded rationality. There is a risk that translating psychiatric phenomena into Bayesian or reinforcement-learning vocabularies imposes theoretical tidiness onto conditions that resist such formalization, potentially obscuring the social, cultural, and developmental factors that shape both pathology and normal judgment. A thoughtful reader might ask whether the clinical laboratory metaphor flatters the model more than it illuminates the patient.
Why it matters for product
The shift from goal-directed to habitual behavior that Palminteri and Wyart identify in addiction maps directly onto how digital products are designed to function: features that reduce friction and automate user choices are, mechanistically, nudges toward habit-system engagement at the expense of deliberative agency. For a CPO, this reframes the ethics of engagement design — metrics like retention and daily active use may be measuring habitual capture rather than genuine value, and product discovery processes that optimize for these signals risk building systems that exploit cognitive vulnerabilities rather than serve user goals. It also raises a structural question: whether product teams should have explicit accountability for auditing when their design decisions shift users from reflective to reflexive behavior.