GPT-4 matches 2,659 human forecasters and improves when averaged with them
RobbWiller · x · 2026-07-23
Human forecast comparison
GPT-4 matched the pooled forecasts of 2,659 people.
Complementarity
Human and GPT-4 predictions were not redundant: averaging them improved accuracy beyond either one alone, reaching r = .89.
What this adds
The figure suggests the model is not only accurate on its own, but also provides information that can complement crowd forecasts.
Related event: Nature Study: GPT-4 Can Predict Social Science Experiment Results(18 posts)→
More from Research
- Meta FAIR's Self-Direction turns agent failures into reusable LoRA skills, +13.8pp on Terminal-Bench — karen_ullrich · 2026-10-27
- DONIC hits 86.79% on bimanual dexterous tracking, crushing prior SOTA's 9.43% — liuziwei7 · 2026-10-27
- Why don't more LLMs ship 10M-token context? A researcher asks about attention tradeoffs — agarwl_ · 2026-10-27
- PSNR below 15 dB in papers likely means the metric was misused, researcher warns — kwangmoo_yi · 2026-10-27
- Ortet's first health eval: 34 open-weight models, 50 tasks, generic rankings fall short — thegautamkamath · 2026-10-27
- Scaling-fit discrepancy: output-head FLOPs are ~65% of compute at the smallest rung, ~3% at 535B — dlwh · 2026-10-27