Needed an application that generates data very fast for a networking application i'm writing but I did notice that /dev/urandom and /dev/random are not very consistent in speed.

Still, i needed something fairly random with more consistent speed. Now, I made an application that caches 1000 randoms upfront and use them for calculation. Now I have my own randomization algorithm backed by the uniqueness of the original rand(). For fun I added data in the set like some phone numbers. I can watch ages to the data to find smth in common or interesting combinations of the data.

I did verify with GPT is the algorithm is unique and it's fail. It generated a complete ML script for itself to check it. Very awesome.

You use urandom, i use retoordom. We are not the same.