OpenAI's AGI economics lead: AI-native internet could be developing countries' mobile-money leapfrog
daveholtz · x · 2026-09-17
David Holtz, a Columbia professor on leave to lead AGI economics research at OpenAI, draws an analogy to how developing countries skipped landlines and traditional banking. In much of the world people still pay for data by the megabyte — for the poorest 40% in Sub-Saharan Africa, 20 GB of mobile data costs about a third of monthly income. AI flips the equation: compute is expensive server-side, but responses can be tiny text payloads, making an 'AI-native internet' of simple text exchanges a plausible leapfrog path.
More from AGI Musings
- Ex-Microsoft exec on AI pause calls: we paused to fix virus bugs without begging Congress — neil_chilson · 2026-09-18
- FOOM is only plausible if P=NP, argues one skeptic of the intelligence explosion — jessi_cata · 2026-09-18
- AI kill switch won't work long-term, says 'Godfather of AI' Hinton — LurkerFromTheVoid · 2026-09-18
- Hollywood fears AI? "Imagine machines recycling the same old stories instead of a room full of executives" — TinfoilTricorn · 2026-09-18
- Z.ai's blog documents early signs of recursive self-improvement in GLM — inductionheads · 2026-09-18
- Gary Marcus: I Warned Four Years Ago About LLMs Hooked to Power Grids — Now We're Here — GaryMarcus · 2026-09-18