The interview takeaway that haunts you
Every interviewer carries one or two interviews that didn't go well, in a way they keep thinking about years later. They're usually the most useful ones.
If you've interviewed enough people, you have one or two you still think about. Sometimes it's the person you said no to who's now a respected leader. Sometimes it's the hire you said yes to who turned out poorly. The pattern is the same: you misread the signal.
The lesson
Calibration is a long-loop skill. You don't know if you made the right call until 18 months later, and by then you've forgotten what your reasoning was. The only way to actually improve is to write down your prediction at the time and review it years later.
What the review looks like
Pull up your old hiring notes. Compare to outcomes. Notice the patterns — the kind of signal you systematically over-weighted, the kind you missed. That's the loop most interviewers never close. The ones who do get noticeably better over a decade.
If your interviewing is no better now than five years ago, you've been practicing, not learning.