It seems the problem logicallee is working is the massively growing destructive power available to individuals or small groups.
Technology is accelerating to the point where the destructive power that was formerly available only to state actors with proper command & control systems is now available to small states, groups, and even individuals -- chemical, bioweapons, delivery by drone, etc. It is now possible to mail-order custom gene sequences for garage bioengineering (yes, they do try to filter the requests against homebrew bioweapons, but the operative word is 'try'). Even computing power -- I'd be surprised if a random dozen people on this forum, properly motivated and funded, could not take down the US power grid within a year.
This scale of mass destruction in the hands of individuals is a far greater scale and scope of problem than the ability of any nutjob to go to WalMart and buy a hunting rifle to point at you, me, or a Congressman.
It is the kind of real problem that keeps serious security pros up at night. And there are many of these scenarios becoming more real all the time, even if logicallee's nuke example seems too fictitious for you.
The real question he's posing is whether its feasible to build an automated system that's sufficiently private and intelligent so that it could scan the comms without violating privacy while only alerting on genuine threats.
I think it's an interesting idea, but even if implementable, would fall to the <Who Guards the Guards?> problem. What is to prevent the people who build, maintain, operate the watch-system from abusing it? Nothing but the same level of ethical training that we have now, so this is simply adding one level of indirection.
But crypto is built from math, which is available to anyone who possesses a brain. Even if you locked down all the academic output related to encryption you can't ensure no one will discover another way to hide and transmit secrets either around or through your usually-private-except-for-serious-threats communications network. You'd have better luck trying to lock down harmful bioagents and fissionable materials, but as you alluded, as long as these technologies exist the world faces a security threat. It seems to me the only way to combat these threats is to construct a society where individuals never feel the need to leverage their increasing power, a police state where anyone's communications can be inspected or abused at will by an agent, a heirarchy, or an algorithm seems antithetical to such a society.
Sure, but that destructive power has nothing to do with online communications. I sort of buy your argument about the power grid, but I am absolutely confident that a dozen motivated and funded people from this forum could easily build our own darknet, operating over some combination of radio waves, dark fiber, and sneakernet, that is completely invisible and unknown to law enforcement.
Again, there have been conspiracies (and armies) in human history for centuries, and most of them didn't have realtime messages in people's pockets. They had letters carried on horseback, and it worked just fine.
People use encrypted chat apps over the internet because it happens to be easy enough and reliably secure enough. If it weren't, there's no inherent reason to keep using the internet for this. There's enough other ways to communicate.
Technology is accelerating to the point where the destructive power that was formerly available only to state actors with proper command & control systems is now available to small states, groups, and even individuals -- chemical, bioweapons, delivery by drone, etc. It is now possible to mail-order custom gene sequences for garage bioengineering (yes, they do try to filter the requests against homebrew bioweapons, but the operative word is 'try'). Even computing power -- I'd be surprised if a random dozen people on this forum, properly motivated and funded, could not take down the US power grid within a year.
This scale of mass destruction in the hands of individuals is a far greater scale and scope of problem than the ability of any nutjob to go to WalMart and buy a hunting rifle to point at you, me, or a Congressman.
It is the kind of real problem that keeps serious security pros up at night. And there are many of these scenarios becoming more real all the time, even if logicallee's nuke example seems too fictitious for you.
The real question he's posing is whether its feasible to build an automated system that's sufficiently private and intelligent so that it could scan the comms without violating privacy while only alerting on genuine threats.
I think it's an interesting idea, but even if implementable, would fall to the <Who Guards the Guards?> problem. What is to prevent the people who build, maintain, operate the watch-system from abusing it? Nothing but the same level of ethical training that we have now, so this is simply adding one level of indirection.