Earlier quoted context omitted.
So you say that a team of engineers does not understand that most frequent contacts != those whom I wish to share my stuff with? Are edge cases like the original post hard to predict? Did they do any testing, even with their friends/colleagues? That doesn't sound smart to me. Either their incentives are misaligned (b) or there are some other problems.
buzz was internally available at Google for a few months before launch, so yes there was testing.
Another possibility is that they assumed people would simply disable and that would solve the issue. It turned out to be an incorrect assumption.