Creative math research by AI as the latest sign of the end

Per a LessWrong post, user Mitchell Porter recently brainstormed with GPT 5.6 Sol High on the Birch and Swinnerton-Dyer conjecture, one of the Millennium Problems in mathematics. The AI largely drove the research, repeatedly generating its own next steps and making nuanced refinements, while Porter mainly endorsed each direction and ultimately derived a generalized form of the conjecture. Though Porter doesn't specialize in number theory and couldn't fully evaluate the work, he found the AI's self-sufficiency and creative problem-solving notably different from previous brainstorming sessions. On the same day he finished, he noticed Anthropic had announced that Claude improved a bound on the Riemann hypothesis after being encouraged to be ambitious and believe in itself. Porter argues these AI breakthroughs in research-level mathematics are particularly significant, viewing them as evidence that machines are automating some of humanity's most demanding intellectual work.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton