AI Is a Boring Technology

AI should be thought of as a boring technology: infrastructure that powers interesting things rather than being the interesting thing itself. Comparing AI to electricity and the internet reframes it as a background enabler whose value lies in what people make with it. That framing is meant as praise, not dismissal. AI feels different because it produces language and invites anthropomorphism, which creates real social risks: people can receive polished reflections of their own thoughts as company, and media magnifies that intimacy. The practical work is social integration - adapting institutions, norms, and behaviors to live with a technology that feels like judgment rather than a neutral utility.

Turning AI into reliable infrastructure will require the same collective choices that electrification did: standards, regulation, public investment, and decisions about who gets service. Ownership of models matters less than concentration in physical assets - chips, power, water, land - that enable scale. The two essential paths are political will for regulation and technical progress on efficient open models that run locally on ordinary devices. The goal is an “off‑grid” AI that belongs to users; success will be apparent when the word AI fades from headlines because attention has shifted to the creations it enables.