This paper shows that code-writing AI agents can take an existing math problem and automatically turn it into a new, harder one while keeping it solvable.
CoDiQ is a recipe for making hard-but-solvable math and coding questions on purpose, and it controls how hard they get while you generate them.