The problem is that, unlike atomic bombs, we can't calculate and make predictions of what will happen, because we would need precisely the same machines that ar

The problem is that, unlike atomic bombs, we can't calculate and make predictions of what will happen, because we would need precisely the same machines that are required to make higher-than-human intelligence. Therefore we don't know, we can't know, so we should at least understand the risks with the best of our very limited understanding, to at least be somewhat prepared when we are able to predict the possible distaster shortly before it happens.