> But no one ever mentions the hundreds of times it quietly wrote code that is better than most engineers can write.

Because the instances of this happening are a) random and b) rarely ever happening ?