DeepMind essay: AGI may arrive as a society of agents, not a single superintelligence
bratton · x · 2026-09-29
A DeepMind Institute essay by Benjamin Bratton, Blaise Agüera y Arcas and James Manyika reframes AGI: since today's best systems are already multi-model teams, AGI may emerge through cooperative societies of agents, tools and humans rather than one winning architecture. The core challenge shifts to orchestrating and governing networks of minds — with intelligence treated as a social phenomenon, echoing how major transitions in life on Earth were highly social.
More from AGI Musings
- Pachocki, Hinton, Bengio & Jack Clark co-author paper on AI R&D automation triggering intelligence explosion — birchlse · 2026-09-29
- AI czar David Sacks slams extinction 'hoax': 4.1% unemployment, doomer job forecasts flopped — DavidSacks · 2026-09-29
- AI doom is the same hysterical grift as campus DEI, argues Kevin Bass — kevinnbass · 2026-09-29
- Bill Gates: AI is an evolutionary event — an intelligent species already smarter than us — haider1 · 2026-09-29
- LLMs are stateless: critic argues the entire AI safety field rests on a fallacy — gerardsans · 2026-09-29
- People use AI to find products but not to buy them: inside the agentic commerce trust gap — iubenda_team · 2026-09-29