This is the second item in this category that Frank Munger, chronicler of nuclear stuff at Oak Ridge, has provided in the last week or so. If you scroll down, I've linked the other in a "Bits and Pieces." This sort of thing happens, probably in Iran as well. I'm recalling another Manhattan Project story, this one at Los Alamos, where someone spilled most of the world's plutonium at that time, in a liquid solution. They ripped up the floorboards and extracted it back out.
It's what happens when you don't entirely understand what you're working with, when you're doing something new. It can be mistaken for carelessness, particularly in hindsight, and that's probably one element, but I would say not the full story. As an experimentalist, I have observed that I have had to learn how to do experiments, and the variables are not always obvious.
Josh Pollack quotes a "technologically sophisticated" person as warning against depending on technical glitches to prevent technological advances. People do indeed rip up the floorboards and extract the plutonium if they have to.
But the glitches do consume time and energy, and they frequently have nothing to do with the science of what is being done, although they can obscure that science. Pile up too many of them at one time, and you can have truly disastrous accidents, or maybe just the destruction of a bank of centrifuges.
This sort of thing is part of why I maintain that it's highly unlikely that non-state actors will ever manage to produce a nuclear weapon. The learning curve is too steep, and there are too many potentially disastrous things that can go wrong along the way.