NBER Paper Explores AI Agent Economics: Plunging Transaction Costs to Reshape Market Design

danielrock · x · 2026-08-10

Daniel Rock argues that "incomplete contracts" is an underrated framework for AI alignment. Trying to specify every rule is doomed to fail, much like the expert systems of the past, primarily because humans struggle to fully articulate their preferences.

Referencing an NBER working paper, The Coasean Singularity?, he highlights how AI agents could transform digital markets. By drastically lowering search, communication, and contracting costs, agents can transact directly to boost efficiency, though they introduce new frictions like congestion. As AI capabilities evolve rapidly, the human inability to articulate preferences has become the primary bottleneck for agent performance.

Original post →

More from AGI Musings

AGI Musings channel →