Small-group debate beats the wisdom of 5,000 individuals, Nature study shows
steverathje2 · x · 2026-11-06
A Nature Human Behaviour study (Navajas et al., 2018) had 5,180 people answer general-knowledge estimation questions individually, then deliberate in groups of five to reach consensus, then revise individually.
Averaging consensus decisions was substantially more accurate than aggregating thousands of independent opinions — as few as four group consensus choices outperformed the wisdom of 5,000+ individuals. The key was reasoning together: structured deliberation within small groups improves collective accuracy rather than undermining it. Relevant to collective decision-making design and, by analogy, debate-style aggregation in multi-agent systems.
More from Research
- Robot success rate doubles 23%→51% from pretraining data alone, no fine-tuning — rohanpaul_ai · 2026-11-06
- Deep-dive report on parallel coding agents: worktree isolation doubles costs, merge conflicts dominate — weswinder · 2026-11-06
- LinQuant: A Quantization Framework for Low-Precision Linear Recurrences in Mamba-2 and Beyond — SonglinYang4 · 2026-11-06
- METR publishes risk assessment on whether AI is accelerating AI R&D — RyanGreenblatt · 2026-11-06
- Shared expert design compared across GLM-5.3, Kimi K3, DeepSeek V4.1 Flash — Nouamanetazi · 2026-11-06
- Kimi K3 swaps sign updates for quantile balancing across its 896 experts — Nouamanetazi · 2026-11-06