Essay Intermediate Argument 90 min read

AI as Normal Technology

Arvind Narayanan & Sayash Kapoor · Knight First Amendment Institute · 2025-04

Narayanan is a professor of computer science at Princeton and directs its Center for Information Technology Policy; Kapoor is a computer science researcher there. Together they wrote the book and newsletter AI Snake Oil, a running audit of inflated AI claims.

Read the original at Knight First Amendment Institute ↗ Open the reading companion

The most influential statement of the deflationary view: AI as a normal — transformative, but normal — technology like electricity or the internet, not a coming superintelligence. Narayanan and Kapoor argue that what matters is not how fast AI capability improves but how slowly adoption and diffusion move through real institutions; that the "capability" measured on benchmarks is very different from reliable competence in messy reality; and that policy should aim at resilience — strengthening the institutions AI lands in — rather than at controlling a hypothetical godlike system.

Read it as the strongest available counterweight to the scenario literature: if the AI 2040 document is the racing worldview taken seriously, this is the worldview in which that race largely fails to materialize. The disagreement between them is the central live dispute in AI politics, and you cannot evaluate either side without the other. Long but cleanly structured; the companion maps which parts to read closely.