People tend to have poor mental models for the long tail of external failures that happen in real life. It's easy to imagine that things that have never happened in the last century would Never Happen. But... they will, somewhere.
Designers are awesome. Sadly they were also unable to find some time in 50 years to raise a wall a few m so it can stand a Tsunami. It seems that the extra scrutiny, was not so extra in the real life when the company will need to allocate real money.
It is trivial to design a system that powers off when unstaffed. Without power, this reactor will SCRAM and passively switch to air cooling over the course of a month.
Presumably, a skilled attacker could compromise the passive safety systems and force a meltdown, but wouldn't it be easier to steal some spent fuel and disperse it?
People are fallible on the best days, assuming everyone did their very best from nuclear physicists to construction workers, mistakes are made. You take steps to reduce the risk. Research gets review. Engineering schematics get review. Construction gets inspection. Still some mistakes will get through.
And people always act their very best all the time right?
You can even have a perfect design, perfect construction, that is mismanaged years after it's built, after the original engineers and bureaucrats lose control.
The same people problems apply to basically every human endeavor, but nuclear's capability to cause accidents that have a lasting impact is pretty scary. You don't feel even a twinge of existential dread when you think about? If you don't, then I don't think I want you working on a reactor.
"Substantially larger" is not the same as "impossible". And, given substantially larger consequences if a reactor pool breaks (compared to a swimming pool breaking), I don't think the question is out of line.
We learned from Fukushima that natural disasters don't always follow the parameters that we expect them to.
