Intelligence got 100x cheaper, so usage exploded 500x: Jevons Paradox in action
luisdans · x · 2026-11-02
Andreas Steno of Steno Research shared a chart illustrating the Jevons Paradox in AI (GPT-4-class intelligence pricing and total consumption, indexed to 2022 = 1, approximate):
- Price side: price per million tokens fell from $20 in 2022 to $2 in 2024 and $0.15 in 2026 — a 99% decline.
- Usage side: total token consumption grew from 1x (2022) to 30x (2024) and 500x (2026).
- Takeaway: intelligence got 100x cheaper, yet total usage grew 500x — a textbook Jevons Paradox, and the author argues "we are still early." Sources: Steno Research, OpenAI/Anthropic price lists, Google & Microsoft disclosures.
More from AGI Musings
- Stripe data: AI adoption predicts pre-pandemic growth too, making AI ROI causally hard to measure — soumitrashukla9 · 2026-11-02
- Early Anthropic employee memoir: co-founder on betting on markets over government AI action — ZeroStateReflex · 2026-11-02
- AI commentator tszzl backs Star Trek's Prime Directive against civilizational flattening — tszzl · 2026-11-02
- The 'thesaurus on steroids': an underexplored interaction pattern for generative models — chrissyykat · 2026-11-02
- rauchg: Own Your Intelligence — Embrace Open Source and Local Models — mitsuhiko · 2026-11-02
- rakyll: No Moat Up, Down, or in the Middle of the AI Stack — rakyll · 2026-11-02