JetBrains found that the popular AI coding optimization reduced output token consumption without hurting performance, but ...
When I watch our trade start handing its tests to language models, I don't feel relief. I feel the same itch I get when a release goes too quiet.