On 16 Jun 2000, lcs Mixmaster Remailer wrote: > > The Foresight Institute is an organization promoting planning for the > impact of future technologies, particularly nanotech. They have a new > set of design guidelines to prevent potential nanotech catastrophes at > http://www.foresight.org/guidelines/current.html. > > One of the points involves encryption: > > > 1. Any self-replicating device which has sufficient onboard > > information to describe its own manufacture should encrypt it such > > that any replication error will randomize its blueprint. > > Anyone have any idea what this means? How would crypto be used here? > > It appears to say that any machine capable of reproducing itself ( an old Artificial Intelligence quest) should use cryptography of some form (checksums, hashes, known-good states etc) so that any error in replication (of itself) would not be further reproduced. Any error in such a device would result not in a device that has it's own blueprint, with an error in it that would be passed to successive generations, but a device with blueprint which is completely random, hence halting the replication process at the first error encountered. At the scale of nano, such a scheme is very efficient, in that the raw materials used up by error-effected devices can easily be re-used. This is an interesting, and effective strategy, especially in a hostile environment, such as the scenario described in Stehpenson's "Diamond Age", where nano devices (nanites?) were constantly warring with each other, polluting the visable world in a kind of 'smart smog.' Since people can't handle macro-tech correctly, (Hello Cherynobyl!) who knows what horrors await humankind in the coming world of nano-tech. spiff
