SMRTR summary
AlphaProof Nexus solved nine decade-old math problems by pairing AI-generated proof steps with a Lean compiler for real-time error feedback—costing just hundreds of dollars per problem, despite a low overall success rate.
SMRTR provides this summary for quick context. The original article belongs to Reddit.
Read the original articleThis archive is built from SMRTR newsletter stories. Subscribe for hand-picked stories without the extra noise.
A small group of mathematicians and developers transformed Lean, a 2013 code-checking tool, into a powerful math proof verifier used by tens of thousands worldwide. By 2025, AI...
Meta and Google's AI systems have tackled century-old mathematical problems, with Meta's model solving stability equations at a 10.1% success rate and Google's Gemini achieving...
Researchers tested AI on unsolved math problems, finding that while advanced systems solved 5-6 out of 10 challenges, AI used brute-force methods rather than elegant proofs...
Axiom, an AI startup, has successfully solved four previously unsolved mathematical problems using its AxiomProver system, including the Chen-Gendron conjecture that had stumped...
DeepMind's AlphaEvolve is a new AI system that addresses problems with machine-gradable solutions, using automatic evaluation to reduce hallucinations. It has shown potential in...
Recent claims about AI solving mathematical problems have generated excitement, particularly ChatGPT's success with some Erdős Problems—over 1,000 mathematical questions from...