Anthropic's secret AI just flirted with a $1M math mystery
An unreleased Anthropic model spent a day and a half chewing on the Riemann hypothesis, math's 150-year-old unsolved puzzle about prime numbers with a $1 million bounty attached. A staffer with zero advanced math training just told the model to try, and it spun up 60 subagents, tested 650 ideas, and pushed the known lower bound of the hypothesis further than ever, confirmed by Anthropic's own mathematicians. It's part of a growing wave of AI-cracked math problems that's got mathematicians publicly fighting over whether AI deserves credit at all.
Great, now even Fields Medalists have an existential crisis to add to their group chat.