[AINews] FrontierCode: Benchmarking for Code Quality over Slop

We made a thing!