Mechanism Realism: Better Architecture, Not Better People

Why do good people inside bad institutions so often produce bad outcomes?This episode introduces mechanism realism: the claim that incentive structures, feedback loops, selection pressures, legal architectures, and coordination protocols determine outcomes more than individual character, intentions, or reasoning quality.The central substitution is simple: stop asking “who decided this?” and ask “what selected for this?” A bureaucracy expands because its mechanism rewards growth. A politician optimizes for the next election because the selection pressure operates on short time horizons. A vague policy word can function as a mechanism by allowing coalitions between incompatible goals.Mechanism realism is not technocracy, libertarianism, determinism, or cynicism. It does not say people are irrelevant. It says individual virtue is not a civilizational strategy. Architecture is the load-bearing variable.The method is: identify the mechanism, distinguish constraints from parameters, predict the output distribution, specify the desired distribution, redesign the parameters, and audit the result.The core claim: outcomes are produced by mechanisms, and mechanisms are engineerable.https://kunnas.com/articles/mechanism-realism