Our hueristic for acceptable risk does not properly account for low-probability but extreme consequence circumstance - proverbial black swan events.

Against instinct, we should find unacceptable even an absurdly minuscule chance of (for instance) irradiating the European peninsula for 10,000 years.

An argument for safety with an attached “even Chernobyl was [mostly safe]” corollary should be suspected of unsupportable optimism.