In the late-80s/early-90s, I used to laugh at the common sci-fi trope of always having some convenient way to hack into anything. I originally assumed this was literary license[3]; surely nobody would actually be dumb enough to to put critical systems on a publicly-accessible network! Well... I guess I was wrong. Not only were we stupid enough to connect everything to a network, some people are trying to start an entire industry based on the idea of plugging things into the internet.
We've simply gotten lucky so far; we have barely scratched the surface of what can be done. I know rare problems are notoriously hard to model. Getting people to properly prepare for these events or even simply understand the actual cost of failure can be incredibly difficult, but that doesn't mean the problem can be ignored.
I almost see a similarity to the martingale[4] betting strategy. It works ok until you get your first rare set of consecutive losses and you bottom out your bankroll. The core problem in martingale is that it requires you to have an unbounded pool of cash to work with, so maybe the answer insecurity is related to only accepting risks if the risk has known bounds.
[1] https://news.ycombinator.com/item?id=10482498
[2] http://www.ranum.com/security/computer_security/editorials/d...
[3] JMS has talked about the need for license when he wrote Babylon 5, where he said that the ships always move "at the speed of plot".
[4] https://en.wikipedia.org/wiki/Martingale_%28betting_system%2...