An economist applies IO theory to AI math: over-entry, efficiency, and tool spillovers all at once
Afinetheorem · x · 2026-09-13
- The author uses free-entry results from industrial organization economics to frame AI math in three ways: (1) if it's 'business stealing' that ignores the fixed cost of training on humans, it resembles over-entry under fixed costs; (2) if it's simply a cheaper way to understand the mathematical world, free entry is generally efficient; (3) if it's a tool with spillover complements that helps humans expand understanding, entry tends to be too low.
- Conclusion: AI math is probably a bit of all three, requiring attention to training costs, diffusion efficiency, and tool spillovers.
Related event: Economists Apply Industrial Organization Theory to AI Diffusion(2 posts)→
More from AGI Musings
- DKokotajlo: real frontier pacing would slow leaders most, not be regulatory capture — connoraxiotes · 2026-09-13
- Opposing AI Slowdown: Telling Small Devs to Stop Is Unethical — TinfoilTricorn · 2026-09-13
- Neel Nanda: current AI progress unsafe without pacing, strike agreements now — NeelNanda5 · 2026-09-13
- Ex-Microsoft vet: 99% of 'don't touch open-source AI' voices just want pricing leverage, like 2001 Linux — aronchick · 2026-09-13
- Cryonics Thought Experiment: Biological Revival Leaves You Eras Behind — matiroy · 2026-09-13
- Telephone Operators' 1920s Lament Is a Perfect Satire of AI Anxiety — CallMePyro · 2026-09-13