Dead code was slowing me down more than bad code

While working in large codebases, I noticed something odd:
bad code was rarely the real blocker — uncertainty was.

Dead code was a big part of that.
Code that still compiled, still passed tests, but nobody knew if it was actually used.

I started experimenting with dead code analysis while building Codur (a VS Code extension for understanding code structure). The goal wasn’t auto-deletion or refactoring — just making unused paths visible so decisions felt safer.

What surprised me was how much mental load dropped even without deleting anything. Just knowing what’s inactive changed how confidently I worked.

Still early, but this reinforced a pattern I keep seeing:
reducing uncertainty often matters more than optimizing code quality.

Curious how others here think about dead code — do you actively track it, or mostly leave it alone?