This seems to conclude that estimates are often wrong (which is true), and then move to the idea that estimates aren't useful. This is wrong. So long as project

This seems to conclude that estimates are often wrong (which is true), and then move to the idea that estimates aren't useful. This is wrong. So long as projects that are estimated to take a week typically take less time that ate estimated to take a month, they're still somewhat useful.