Recent OpenAI releases of hundreds of AI-generated proofs, including a claimed resolution of the rational Hodge conjecture for CM abelian varieties, have elevated that problem’s standing in trader sentiment. This partial algebraic geometry advance, verified in some cases via Lean formalizations and building on the company’s earlier Navier-Stokes claim, signals stronger near-term AI traction there than for number-theoretic or complexity-focused challenges like the Riemann hypothesis or P versus NP. The near-tie with “no solution by December 31, 2027” underscores verification hurdles, community acceptance requirements, and typical multi-year timelines even for AI-assisted results, while Birch and Swinnerton-Dyer sits close behind amid similar but less publicized progress. These dynamics highlight how demonstrated model capabilities in specific mathematical domains are shifting implied probabilities.