Google DeepMind’s AlphaProof Nexus Autonomously Demolishes Nine Open Erdős Mathematical Problems
AI is shifting from a basic text predictor into a logical thinker by mastering formal mathematical frameworks. On May 27, 2026, scientific analysis co
AI is shifting from a basic text predictor into a logical thinker by mastering formal mathematical frameworks. On May 27, 2026, scientific analysis confirmed that Google DeepMind’s unreleased model family, AlphaProof Nexus, autonomously resolved nine out of 353 open Erdős problems, with two specific algebraic geometries remaining completely uncracked by human scholars for 56 years. The technical breakthrough was achieved by pairing large language models with the Lean formal proofchecking assistant, allowing the neural engine to continuously verify every sequential calculation step for just a few hundred dollars in cloud resources. Academic computer research is entering an era of "Autonomous Logic Discovery," where systems evolve past processing historical human records to independently construct verifiable algebraic theories. Advanced software verification teams are already tracking these automated proof setups, utilizing the mathematical code to stresstest highsecurity smart contra
Read More..
B2B Tech News | 2 days ago