Random Number Generation

27 September 2024

The weirdness of random number generation in computation, given how often it’s used in computational art, game design, cryptography, and more, has become increasingly fascinating to me—particularly the idea that being able to identify or develop the underlying algorithm that generates random numbers nullifies their true randomness. If you can describe how it was created, it’s not exactly “random.”

I hadn’t done much exploring beyond this fact before today, when I discovered the existence of hardware random number generators that sample the natural entropy of things like atmospheric noise, circuitry, and light to produce their values.

I find this absolutely delightful. There’s something metaphorically quite rich about the fact that this feature so commonly implemented in computational practice can only truly exist in the wilds of the material world; in the tightly-controlled reality of the computer, it can only exist as an approximation without some sort of outreach across worlds.

It makes me think that a hypothetical civilization of wholly computational beings would probably end up developing something like a theology around randomness.