30. Underestimate of precritical events as elements of coming global catastrophe

          If as a result of some events, the probability of global catastrophe substantially increases (for instance, we find an extremely cheap way of enriching uranium), that event may contribute a large amount of probability mass to the likelihood of global disaster, but be underweighted in importance because it is not a disaster itself. Another example is that of a nuclear war--although a nuclear might kill “only” one billion people, the ensuing nuclear winter and breakdown of polite civilization would be much more likely to usher in human extinction over the long-term than if the war never occurred. We call these “precritical events”. More effort needs to be devoted to dangerous precritical events, which may kill no one in and of themselves, but contribute substantial probability mass to global catastrophe.