DeepMind's 'What comes after AGI?' maps four paths to ASI
AGI Hunt · wechat · 2026-11-20
DeepMind co-founder and chief AGI scientist Shane Legg shared a new DeepMind Institute article "What comes after AGI?", adapted from Tim Genewein's arXiv report "From AGI to ASI" (14 authors including Marcus Hutter, Allan Dafoe, and Legg).
The piece frames intelligence as a continuum, defining ASI as consistently beating large teams of human experts on virtually all tasks with fewer resources. It introduces "cognitive industrialization": global effective compute grows 10x per year (hardware cost curves, 3-6x/year algorithmic efficiency, rising investment), so cognitive work costing $1B in compute today could cost $1,000 in a decade. Four paths to ASI are mapped: scaling models and data (data walls, diminishing returns, physical limits), paradigm shifts (persistent learning and sample efficiency gaps), large-scale agent collaboration (needs multi-agent scaling laws), and recursive self-improvement (fastest in auto-verifiable domains like math and code). Compute limits alone can't slow multi-pathway progress and may create capability overhangs; the answer is rigorous model-measure-forecast tracking.
Related event: DeepMind Maps Paths from AGI to Superintelligence in New Paper(4 posts)→
More from AGI Musings
- Bo Dai and Yilun Du release EBM tutorial unifying ML through one energy-based lens — daibond_alpha · 2026-11-20
- Lex Sokolin podcast: you can beat 98% of competitors and still get wiped out by the wave you missed — LexSokolin · 2026-11-20
- Investor bullish on Anthropic stock as one of few companies that reach RSI — _arohan_ · 2026-11-20
- Economist Daniel Susskind brings sensible AI talk to The Daily Show — joshgans · 2026-11-20
- Denny Zhou: since 2026, the best AI models make the best data — denny_zhou · 2026-11-20
- Do LLMs kill off-computer thinking? Two researchers disagree — Afinetheorem · 2026-11-20