they are optimizing against false positives, and find false negatives a tolerable loss.

thats the rationale to make it not crazy, whether this is the best way to accomplish that is not resolved.

“Cracking the Coding Interview” covers this pretty early in the book