WORK
A Conflict of Visions
Thomas Sowell's A Conflict of Visions (1987) proposed that most political and social disagreements are not about specific policies or data but about fundamentally different assumptions regarding human nature and the possibility of social improvement. The constrained vision sees human beings as fundamentally limited—selfish, shortsighted, prone to error—requiring institutions that channel flawed nature toward tolerable outcomes. The unconstrained vision sees human nature as malleable and improvable, with limits representing temporary obstacles that reason, effort, and better design can overcome. The book examined how these visions shaped debates on crime, war, equality, and governance across centuries. Sowell did not argue for one vision over the other but demonstrated that they produce different questions, different standards of evidence, and different criteria for acceptable solutions.
In The You On AI Field Guide
Published nearly four decades before large language models arrived, A Conflict of Visions supplied the analytical framework that makes the AI discourse intelligible as something other than random disagreement. Sowell traced the constrained vision through Adam Smith's invisible hand, Edmund