21Mar
Random based on video input, why isn’t there some predictability [migrated]
I just saw this article and many like them before where they use a video of some random movement and a hashing function to make a cryptographic random number generator
but I'm always curious why this is as "unpredictable" as they claim. These random generators, they're supposed to help with random number generation with the assumption you can't predict what the next frame will look like.
But if you know the hashing algorithm they use and have frame 1, you can't predict frame 2 with a 100% accuracy, I admit, but wouldn't it still greatly reduce the search space because you can assume that the pendulums for example follow the laws of physics?
