[natmaka]

> There are some very recent arsenic spills

Indeed, but nothing comparable to the spills at Chernobyl or Fukushima.

> Nuclear fuel can be recycled

Only once, and France, an industrial leader in this area, only manages to recycle 10% of its reactor fuel this way.

> It's mostly not done because it's cheaper not to, just like in renewables

No, that's completely false. Closing the fuel cycle was considered the Holy Grail as early as the 1950s, because everyone knew that uranium deposits would greatly limit the expansion of nuclear power. The industrialization of reactors of the most promising architecture (fast neutron-breeders, sodium-cooled) as well as others, attempted at great length and expense in many countries, failed everywhere ( https://en.wikipedia.org/wiki/Breeder_reactor#Notable_reacto... ), there is no ready-to-deploy model of such reactor, and this approach has been virtually abandoned, replaced by the pursuit of fusion.

> The danger was known based on multiple data points

Before the major nuclear accident at Fukushima, the formulas for calculating seismic risk (the tsunami was triggered by an earthquake) were incorrect because they neglected very old earthquakes. The cause was an inability to properly assess the risk. This inability was not universal, as some (for example, Y. Hirai in the case of the Onagawa nuclear power plant, which was closer to the earthquake's epicenter and withstood the earthquake: https://en.wikipedia.org/wiki/Onagawa_Nuclear_Power_Plant#20...) did take the necessary precautions.

> They acknowledged evacuation was not necessary in the way it was implemented.

Arms-chair tacticians are verbose after the fact, but nowhere to be found when the problem was an ongoing challenge and experts described the worst-case scenario. The testimony of the prime minister at the time, referenced above, is perfectly clear.

> Capacity factor is important to understand how much firming you need.

No value is prohibitive, as there are many other pertinent parameters.