Is there any crypto strong enough to withstand the analysis of a weakly godlike agency?
Just a fun thought experiment.
Since we don't know how many giga^H^H^H^Hgoogol-flops they have to work with, we might expect that they can read a message that Alice sent Bob by getting to know Alice, getting to know Bob, then resimulating their consciousnesses and then reading the message traffic the resimulations send each other.
Or, I don't know, maybe they know how to quickly factor every number at once (which probably makes public-key anything completely worthless). For that matter, maybe they've already installed a rootkit on Physics 1.0 and can read our one-time-pads without collapsing their wavefunctions.
So, make certain assumptions about their capabilities:
* They are limited to the power output of a single star. (Charlie Stross, lag-bandwidth requirements)
* If they use reversible computing, their memory capacity upper limit will be detectable via gravity. (deletion: waste heat)
* If they don't, they'll pump out IR like nobody's business. (Larry Niven, Dyson Spheres, big brown dwarves, information entropy)
* Their googolflops+bandwidth won't exceed the Bekenstein-Hawking limit *or* the speed of light.
* They've probably turned their planets into computronium, but they still need their sun. (total conversion, baryon number)
These assumptions are based on relatively uncontroversial physics. And it's probably enough to get a quick back-of-the-envelope guess as to their power. Of course, we can assume they are better at math than we are. Much much much better... so we stick to crypto that is provably strong. One time pads only work if you sometimes have a secure channel. I'm not sure we do. Quantum encryption, now, maybe that will hold up. Don't understand it, so who knows?
Since we don't know how many giga^H^H^H^Hgoogol-flops they have to work with, we might expect that they can read a message that Alice sent Bob by getting to know Alice, getting to know Bob, then resimulating their consciousnesses and then reading the message traffic the resimulations send each other.
Or, I don't know, maybe they know how to quickly factor every number at once (which probably makes public-key anything completely worthless). For that matter, maybe they've already installed a rootkit on Physics 1.0 and can read our one-time-pads without collapsing their wavefunctions.
So, make certain assumptions about their capabilities:
* They are limited to the power output of a single star. (Charlie Stross, lag-bandwidth requirements)
* If they use reversible computing, their memory capacity upper limit will be detectable via gravity. (deletion: waste heat)
* If they don't, they'll pump out IR like nobody's business. (Larry Niven, Dyson Spheres, big brown dwarves, information entropy)
* Their googolflops+bandwidth won't exceed the Bekenstein-Hawking limit *or* the speed of light.
* They've probably turned their planets into computronium, but they still need their sun. (total conversion, baryon number)
These assumptions are based on relatively uncontroversial physics. And it's probably enough to get a quick back-of-the-envelope guess as to their power. Of course, we can assume they are better at math than we are. Much much much better... so we stick to crypto that is provably strong. One time pads only work if you sometimes have a secure channel. I'm not sure we do. Quantum encryption, now, maybe that will hold up. Don't understand it, so who knows?