The Economic Impacts of AI: A Multidisciplinary, Multibook Review [pdf] (kevinbryanecon.com)

Kevin A. Bryan’s essay reviews seven books by social scientists on AI’s economic effects and finds a mixed but valuable picture: economists have developed robust frameworks for integrating AI into existing production functions—most notably the “AI as cheap prediction” framing, the economics of data as an asset, and clear analyses of who benefits—but these books offer limited guidance on the large, transformative scenarios many AI researchers worry about. Bryan surveys works ranging from The Second Machine Age and Prediction Machines to more recent treatments of data economics and organizational implementation (e.g., The Skill Code, Co-Intelligence) and a Silicon Valley–style speculative treatise (Situational Awareness). He highlights that books play a unique role in synthesizing fast-moving research into accessible policy-ready narratives even as the underlying academic literature—NBER papers on AI and LLMs surged after 2017, with over 130 LLM papers and 103 on generative AI by May 2025—rapidly expands. Technically, Bryan contextualizes why Silicon Valley anticipates epochal change: deep learning’s hierarchical feature learning, gradient-descent optimization, ReLU activations (which mitigated vanishing gradients), and the confluence of massive digitized data and GPU compute enabled modern LLMs and generative models. The review stresses practical barriers—implementation, firm architecture, and sociological factors—that temper rapid adoption, but also flags critical gaps: economists need to produce rigorous, actionable advice on policy responses to potential rapid labor churn, macroeconomic shocks, accelerated scientific progress, capital–labor shifts, and even existential risks tied to advanced AI.