This paper extends the Cecchetti-Kharroubi (2015) talent misallocation framework from the financial sector to global software engineering. Using developer population data from SlashData, Evans Data Corporation, and JetBrains, combined with AI code generation statistics from GitHub and GitClear’s 153-million-line code analysis, I estimate that approximately 97% of the world’s 47 million software developers perform structurally automatable template-composition work—a misallocation conservatively valued at 0. 7–1. 4 trillion annually. The paper introduces two novel contributions: (1) the Mode A/Mode B distinction between cost reduction and function elimination in labor displacement, arguing that software engineering faces the latter; and (2) five structural mechanisms explaining why this misallocation persists despite its scale, integrating bounded rationality (Simon, 1955; Kahneman, 2011) with Adaptive Compression Advantage Theory (ACAT). Policy implications for educational investment, workforce reallocation, and corporate governance are discussed.
Ryoji MURATA (Wed,) studied this question.