You need a sufficiently large reserve to allow the reactor to cool, not an infinite supply. Reactors can be shut down and in this case the pool is sized to absorb all decay heat from the shutdown, plus a significant safety margin.
Depending on a lack of incompetence in dangerous systems works until it doesn't. To the extent that these things can simply halt when incompetently managed, they should.
If anyone disagrees, I'd like to know why they think the next hundred years are going to be so much freer of political shortsightedness and corruption than the last hundred.
Sounds like a quite simple design indeed, and only one valve opening away from use?
Engineering a very resilient pond does not feel like as complex a problem as engineering highly complex cooling systems to be resilient.
2. There have been proposals before for core dilution buckets: a wide shallow dish under the reactor full of something like gallium for the hot core to dissolve into. As it spreads out into the dish, the heat and radiation fluxes become less unmanageable, and the core material becomes less critical.
3. Don’t let that happen. :-P