Comment Re:Can someone help explain "perfect" randomness? (Score 1) 140
In practice multiple sources of random numbers are used and the sources need to be trusted using a chain of trust. To deal with micro bias two streams from the same source can be merged with one inverted. All this depends on how secure you want to be. Applications typically just need a random seed and then use a PRNG to generated subsequent values. High valued long term keys need to be held in redundant hardware vaults. There are so many ways into systems that, unlike movies, breaking encryption is almost pointless.