I Pointed My AI Research Engine at Goldbach's Conjecture — It Found a Hidden Bias (2026)

Short summary

Developer built Luka, an AI research engine, and ran it against Goldbach's conjecture—discovering a previously unmeasured Chebyshev bias in prime-pair distributions. Analysis of 2.5M Goldbach counts revealed n≡1 (mod 3) has 0.26% more representations than n≡2 (mod 3), with statistical significance p=4.07×10⁻²⁰⁴. The bias stems from twin-prime preferences and demonstrates AI's capability for autonomous mathematical discovery.

• •AI research engine (Luka) discovered hidden Chebyshev bias in Goldbach's conjecture analysis
• •Analyzed 2.5M even integers with statistical significance p=4.07×10⁻²⁰⁴, contradicting Hardy–Littlewood formula
• •Demonstrates autonomous AI systems can make rigorous mathematical discoveries

Generated with AI, which can make mistakes.