Who says that AI systems don't also destroy themselves during takeover, killing all humans in the process?
a Schelling point for those who seek one
