DeepMind essay: agent coordination, not solo competence, will flip the economy
krishnan · x · 2026-10-30
Most agent evals score one agent working alone. A new Google DeepMind essay by Shane Legg, Alex Imas and Andrew Koh argues the real economic shift will come from what those evals skip: how well agents work together.
- They split AI progress into competence and coordination: competence moves wages and prices gradually; in their model, coordination past a threshold flips returns to scale, letting large AI-run firms suddenly outcompete smaller ones.
- In their data, people paired to complement each other beat random pairs by 32%.
- Their own admission: today's AI systems coordinate poorly in practice.
- The benchmark checklist enterprises should steal: split work without duplication, pass context cleanly across handoffs, catch each other's mistakes instead of compounding them.
Related event: DeepMind paper: coordination, not solo agents, drives economic change(2 posts)→
More from AGI Musings
- ML veteran Andriy Burkov calls 'money won't matter post-superintelligence' nonsense — burkov · 2026-10-30
- 'My job is now arguing with computers' — and it won't end anytime soon — danfaggella · 2026-10-30
- Kevin Roose's Machine Gods pod: why scientists are mad at OpenAI solving their hardest problems — kevinroose · 2026-10-30
- Ethan Mollick on AI and work: 'You're a hand-weaver in a world of power looms' — emollick · 2026-10-30
- Vercel founder Rauch: trust is the deepest moat for a post-AGI business — soleio · 2026-10-30
- Scattered Across AI Assistants: A New Kind of Cognitive Failure Nobody Has a Prior For — manosaie · 2026-10-30