Also, complexity in and of itself can be handled. The issue is that handling it requires discipline, and for civilian nuclear power in the US, the discipline has often not been there. (By contrast, military nuclear power in the US has always had strong discipline, and as a result relatively complex reactors are run safely.)
No, they thought they could site the backup diesel generators and switchgear behind a seawall that turned out not to be high enough when a tsunami came through. None of the issues at Fukushima had anything to do with the design of the reactor itself; in fact, the reactor was obviously significantly over-engineered given how well it withstood an extended loss of backup cooling.
Not if the public's reaction is "Aah, nuclear bad, no more nuclear!" instead of "hey, idiots, you need to change your specs for seawalls".
> I'm sure they thought the seawall was high enough when they built it.
The problem was that their definition of "high enough" was wrong. That's not a matter of reactor design, or even an issue that's specific to reactor design; it comes up in any issue involving cities or facilities near seacoasts (for example, specs for levee heights). But in every other area that's affected by this issue, the response is "you need to change the specs for how severe an event the facility needs to withstand", not "stop building this facility altogether". Nobody argues for abandonment of all coastal cities because some levees aren't high enough.
> Why are we now so sure we couldn't make a similar miscalculation?
Who said we were? There is no such thing as zero risk.
Elon uses high PITA factor when talking about why SpaceX isn't pursuing hydrogen as a rocket fuel. It similarly looks awesome on paper but results in higher TCO and higher complexity in practice. It's cheaper (according to SpaceX) to use kerosene or methane and build a bigger rocket.
For nukes vs. renewables it's likely overall cheaper to just over-build renewables and storage than to handle the long tail of pain from current-type nuclear. Spend a few billion dealing with nuclear waste, or spend a few billion on more generation and storage capacity. The former is sticky and complicated, while the latter is just straightforward commodity scaling of stuff everyone already knows how to build and handle. It's stupid and simple and scalable and if it's not enough just do more of it. In other words: just build a bigger rocket.
What I say applies to conventional and near-term fission. Fusion and other exotics might be a different ballgame. Fusion would likely be complex and high capital cost but it doesn't have the waste and "black swan failure" issues that pee in the pool for fission.
I also think PITA factor is kind of a death by a thousand cuts -- a "long tail" as I said.