>"how do we know? Surely the smart ones aren't getting caught so we never hear about them."
This only works if you exclusively consider crimes that have zero discernible external effect ("victimless crimes" in common parlance). Some percentage of copyright infringement or drug use would be good examples. But most crimes people really care about have external effects, and that is indeed precisely why they're crimes. If someone robs a house, they need not ever be caught for everyone to know a house was robbed. In turn, one can consider the ratio of crimes to criminals-caught-for-those-crimes. Further, since trials are all fully public, one can look at why exactly those criminals were caught. If most get caught, and it tends to be for stupid mistakes, then one can reasonably argue "criminals aren't very smart".
Leaks by definition are self-revealing, that's the point of them. If secret information is appearing in the general world, the organization knows their are one or more leaks, even if the exact nature of them could be a number of things (ie, could be an insider, could be a partner, could be an external break-in).
To a degree, but just because a crime was noticed doesn't mean that that the culprit was found. There's an awful lot of unsolved crimes out there.
But I also remember what a police officer told me once -- that unless you're a total idiot, you can get away with almost any crime, provided that you only do it once and never break any law again. Most criminals are caught because they do it repeatedly.
Personally, I do think that most criminals are dumb, based on the fact that if you're smart, you can make more money in the long run through legal means.
This only works if you exclusively consider crimes that have zero discernible external effect ("victimless crimes" in common parlance). Some percentage of copyright infringement or drug use would be good examples. But most crimes people really care about have external effects, and that is indeed precisely why they're crimes. If someone robs a house, they need not ever be caught for everyone to know a house was robbed. In turn, one can consider the ratio of crimes to criminals-caught-for-those-crimes. Further, since trials are all fully public, one can look at why exactly those criminals were caught. If most get caught, and it tends to be for stupid mistakes, then one can reasonably argue "criminals aren't very smart".
Leaks by definition are self-revealing, that's the point of them. If secret information is appearing in the general world, the organization knows their are one or more leaks, even if the exact nature of them could be a number of things (ie, could be an insider, could be a partner, could be an external break-in).