On Oct. 29, 2026, Google DeepMind’s @alexolegimas announced The Economics of Intelligence, which he said he released with Andrew Koh and Shane Legg. The essay examines when advances in AI might affect wages, employment and market structure.
The difference between skill and teamwork
The authors distinguish competence, an agent’s ability to perform individual tasks such as writing or coding, from coordination, its ability to work with other agents inside a firm. In their framework, better individual skills gradually change wages, prices and the mix of human and AI workers while leaving industry structure largely intact. Better coordination, if it crosses a critical threshold, produces abrupt shifts in market concentration, employment and wages.
The distinction matters because, the essay argues, an AI agent that outperforms people one-on-one may still lose to a well-organized human team. To illustrate the value of complementary skills, the authors describe a quiz given to 100 people across six domains. They report that pairs matched to complement each other outperformed random pairs by 32% on a task. That comparison concerns the pairs in their exercise, not a measured advantage for AI teams.
In the essay’s minimal AGI scenario, the authors project that workers whose skills AI can match would see their wages decline toward the cost of computing, while specialists and human coordinators could earn more. @alexolegimas predicts AI’s impact will likely appear in wages before employment or market structure changes. The essay describes possible capability regimes and says it draws on a separate paper, Organizational Intelligence, for the full economic model; its projected shifts are not observed economy-wide results.
Comments