This is a semantical argument. The software worked just fine, just the human operating it made a procedural mistake.
I'm not an election expert, but if you read the article below, they figured the error out because the printed count and the software count skewed due to human error were different. This can easily be checked. So other counties would notice a discrepancy if it occurred. Even if I were to accept your argument that a bunch of small errors went unnoticed, it wouldn't make much of a difference anyway.
https://www.clickondetroit.com/decision-2020/2020/11/05/why-is-antrim-countys-vote-count-at-0/