It is relevant because if we are dealing with a poisson-like distribution the chances of the event happening is greater over a longer time-span. We don’t know w

It is relevant because if we are dealing with a poisson-like distribution the chances of the event happening is greater over a longer time-span. We don’t know what distribution nuclear strikes and accidents occur with. It may very well be a distribution similar to Russian roulette. If it is, we really really shouldn’t risk it any further. That is the flaw in your logic. It is assuming a favorable distribution when the distribution is unknown.