CONCEPT
Verification Literacy
The ability to read and evaluate code — to trace its logic, identify its assumptions, and determine where it will fail — even if one cannot write it. The specific form of competence the AI era requires and few curricula teach.
Verification literacy is the form of programming competence that the AI era has made structurally necessary. It is not the ability to write code, which AI tools have increasingly made optional. It is the ability to read and evaluate code — to trace its logic, identify its implicit assumptions, and determine the conditions under which it will fail. The distinction matters because the two skills, while related, are not identical. A builder who cannot write code may nonetheless learn to read it critically, and the reading skill is precisely what makes the difference between deploying AI output responsibly and deploying it as a bet. The concept is implicit throughout Dijkstra's writing — particularly the 1988 essay on teaching computing — and has become explicit in the pedagogical literature that has emerged in response to the ubiquity of AI code generation.
In The You On AI Field Guide
The split between writing and reading