0
Applied AI·July 20, 2026·1 min read

AI just disproved a major math problem. A mathematician explains.

Share

If a system like Anthropic’s Fable 5 can contribute to disproving a major open problem, the bar for “AI as research collaborator” just moved from toy examples to frontier math. R&D-heavy orgs should be running small, focused pilots where models are tasked with conjectures, counterexamples, and proof sketches—not just literature summaries.