Amdahl's law rebuts takeoff skepticism: last few percent of AI automation delivers most gains
austinc3301 · x · 2026-11-09
Responding to Ramez Zia's claim that fast-takeoff models use gentler diminishing returns than empirical data show, austinc3301 argues the core flaw is treating research productivity per capability point as constant. Invoking Amdahl's law, he notes it depends on the share of research AI can automate: the larger the fraction, the higher the maximum speedup, with the last few percent of automation delivering most of the productivity gain. The AI Futures Model assumes the feedback loop is weak today but strengthens over time, whereas Ramez treats loop strength as invariant.
More from AGI Musings
- Everyone is automating white collar work — the real moat is blue collar — JoeJustice · 2026-11-09
- Cathie Wood: AI Costs Fall 75-99% a Year, CPI Badly Undercounts Deflation — rohanpaul_ai · 2026-11-09
- The models *are* the product: just add enough structure without over-constraining — ivan_bezdomny · 2026-11-09
- 'You Can Get Arbitrarily Smart Without Reading Books' — But Does Anyone Have an Example? — alexisgallagher · 2026-11-09
- Future CEO of agentic orgs will be a system designer, not a McKinsey alum — GregCook2011 · 2026-11-09
- Box CEO: agent swarms turn AI productivity gains from 20-50% into unbounded — inductionheads · 2026-11-09