Rereading Bostrom's Superintelligence: 2014 predictions all arrived early
kimmonismus · x · 2026-11-01
Rereading Nick Bostrom's 2014 Superintelligence on flights, the author finds its predictions aged remarkably well — and arrived far faster than expected.
- Bostrom estimated Go programs were at strong amateur level in 2012 and might beat the world champion in about a decade; AlphaGo beat Lee Sedol in 2016, and AlphaGo Zero mastered Go from scratch in 2017.
- The book flagged visual scene analysis, object recognition and robot control as hard, and framed human-level natural language understanding as "AI-complete" — by 2026 nearly all of it has materialized, ahead of schedule.
- The author claims the hardest unsolved math problems are now being solved weekly and recommends rereading the book to grasp how exponential current progress is.
More from AGI Musings
- Google's 'Goodnight' datacenter in Claude, Texas projected to host 2.27M H100-equivalents by late 2027 — scaling01 · 2026-11-01
- Will big labs really refuse to sell on-prem year-old models? Sovereign AI bets questioned — Afinetheorem · 2026-11-01
- davidad podcast on emergent wisdom in LLMs, says Anthropic should stop cyber-offense training — RichardMCNgo · 2026-11-01
- AI commentator: flashing credentials to pass off contested views as consensus erodes trust — AndyMasley · 2026-11-01
- FT data: US AI sector productivity surges while only Sweden keeps pace — Zachly · 2026-11-01
- Kenneth Stanley: the thousands of math problems AI fails to solve are an underrated scientific goldmine — kenneth0stanley · 2026-11-01