Evolution halts at the point where a single individual dooms the entire population. It cannot work with a step fitness function, with no margin for error.
And by evolution I mean the abstract understanding for evolution that we have, which is the core of the genetic algorithms (not just biological evolution, nor some abstract "a population getting better in something").
I think we are encountering the problem at different scale at first: some countries can destroy the entire species, then more and more countries do, then it would become possible for some large or committed enough organizations. We are learning to cope along the way, and the approach so far is to keep knowledge under covers, and not to give such powers lightly.
And by evolution I mean the abstract understanding for evolution that we have, which is the core of the genetic algorithms (not just biological evolution, nor some abstract "a population getting better in something").
I think we are encountering the problem at different scale at first: some countries can destroy the entire species, then more and more countries do, then it would become possible for some large or committed enough organizations. We are learning to cope along the way, and the approach so far is to keep knowledge under covers, and not to give such powers lightly.