Disturbingly Flawed Interview Processes — A Cri de Coeur
The research is not ambiguous. Unstructured interviews perform worse than random selection. Behavioral interviews measure impression management, not competence. LeetCode exercises test rote memorization in a context disconnected from the work. The methods that would work are known — and have been for decades. The industry has chosen not to adopt them. This post examines why.