FAR framework enables AI to solve 1977 Erdős–Straus conjecture
wellecks · x · 2026-08-19
Core Concept
Researchers introduced FAR (Find, Attempt, Recommend), exploring if AI can enable a more scalable mode of mathematical research rather than just accelerating existing processes. Instead of starting from pre-selected problems, it begins with a research direction, searches literature for open problems, attempts them at scale, and filters results for expert review.
Key Results
- One pilot run over combinatorics literature surfaced potential resolutions to hundreds of open problems.
- Included an answer to a 1977 Erdős–Straus question and a counterexample to a conjecture cited as open in a 2025 survey by Tao.
- None of these problems were selected in advance, showcasing AI's ability to autonomously discover and solve problems.
More from AGI Musings
- AI Era Decision Surge: Brains Must Handle More Judgments — bennash · 2026-08-19
- Linear Designer: AI forces unnatural verbalization of intuition — round · 2026-08-19
- Silicon Valley AI bubble widens gap with public perception — nickbaumann_ · 2026-08-19
- Opinion: AI Will Hypercharge the K-Shaped Economy — scaling01 · 2026-08-19
- Researcher warns AI 'slop explosion' is breaking peer review and science — ParshinShojaee · 2026-08-19
- LLMs as the Internet's Immune System: Smoothing Anomalies and Stifling Innovation — tom_doerr · 2026-08-19