GPT-5.6 Solves Stats Problem With Counterexample
garrytan · x · 2026-07-15
This post highlights that GPT-5.6 solved a long-standing open problem in statistics by providing a single counterexample, with the related paper having amassed over 130,000 citations.
The original text outlines the background:
- The research focuses on FDR control in multiple hypothesis testing.
- The classic Benjamini-Hochberg method is only proven valid when data points are independent.
- However, much of the real-world data we deal with is correlated.
Commenters find this particularly impressive because:
- The model genuinely helped advance a foundational statistical issue.
- It demonstrates how the pace of scientific exchange on X is accelerating.
- It serves as yet another example of scientific conclusions being updated through the discovery of counterexamples.
Related event: GPT-5.6 Reportedly Found a BH Counterexample(8 posts)→
More from Models
- Kimi K3.1 Pro tops MathArena open models, closing frontier gap to about 6 months — teortaxesTex · 2026-10-26
- Ruff creator Charlie Marsh: frontier models find infinite bugs in any software — charliermarsh · 2026-10-26
- Dev testing one skill across Opus 5.5 and Sol 6.1 finds instruction following 'all over the map' — lucasmeijer · 2026-10-26
- supermemory unveils learner-1, claiming continual learning for any model — julianweisser · 2026-10-26
- Anthropic says Opus 5.5 cuts cost per task by ~40% vs Opus 5 — taufiqintech · 2026-10-26
- AI search APIs ranked: top 12 within 7 points, yet prices differ nearly 7x — FinanceYF5 · 2026-10-26