Now let's talk about costs. Not idealized costs, but actual, non-hidden costs of building, maintaining, and decommissioning nuclear power plants (oh and storing the waste in effective perpetuity).
What's that you think decommissioning would be too expensive and want to keep running the plant extra long with upgrades? Oooops, that's how you get Fukushima.
The only reason we aren't upcycling the 'waste' is because of an order issued by Jimmy Carter to prohibit the proliferation of Plutonium; we could be using it as fuel in other reactors.
This is why the real lesson is to build defenses in depth, with enough variety to protect against common mode failure. Fukushima Daini - only 12km away from Fukushima Daiichi - suffered similarly from the earthquake and tsunami, but a single surviving pump (and the skill and effort of the staff) was enough to prevent Daini from suffering the same fate as it's sister powerplant.
Good engineering asks questions like, "Are we safe when (not "if"!) the entire bank of backup generators fail?", or "Can a single event cause catastrophic failure?"