Announcements over recent months that artificial intelligence models have cracked previously unsolvable problems are shaking top-level mathematicians to the core, and they are now asking themselves: Where do we come in?