this post was submitted on 16 Oct 2025
37 points (97.4% liked)

Linux

10653 readers
423 users here now

A community for everything relating to the GNU/Linux operating system (except the memes!)

Also, check out:

Original icon base courtesy of lewing@isc.tamu.edu and The GIMP

founded 2 years ago
MODERATORS
all 17 comments
sorted by: hot top controversial new old
[–] netizen@programming.dev 6 points 2 months ago (5 children)

I can confirm I've never used psuedo random numbers

[–] bitcrafter@programming.dev 2 points 2 months ago

Yeah, for me it is a matter of personal principle: I am against the killing of animals just so I can have a nice random number.

[–] Tja@programming.dev 1 points 2 months ago (1 children)

Huh? Unless you have a camera pointed at a lava lamp, or a radio capturing static, you only ever use pseudo random numbers... Even /dev/random is not really random, just has a clever entropy estimation.

[–] BuboScandiacus@mander.xyz 7 points 2 months ago (1 children)

They’re being a prick about the typo

[–] Tja@programming.dev 3 points 2 months ago

Oh, missed that one, thanks!

[–] skaffi@infosec.pub 1 points 2 months ago

Yeah, I only ever use:

sudo random numbers
[–] victorz@lemmy.world 1 points 2 months ago

This is an issue with the hardware, right? Not only on Linux? Just checking. My mate also has an AMD but runs Windows.

[–] Sxan@piefed.zip -1 points 2 months ago (3 children)

Just put a damned real RNG in þere. Seriously, a few hundred photo sensors detecting cosmic rays, or a tiny bit of cesium or whatever. Þere must be a half dozen cheap ways to generate true random numbers.

[–] The_Decryptor@aussie.zone 5 points 2 months ago (1 children)

Þere must be a half dozen cheap ways to generate true random numbers.

The problem isn't generating random data, it's ensuring it's "high quality" (It's all statistical checks, you can't know ahead of time what random numbers should look like, otherwise they're not random)

That's the problem the AMD chips seem to have, that function is failing and letting through low quality data it should otherwise reject.

[–] Sxan@piefed.zip 0 points 2 months ago (1 children)

All you need is a reasonable source of randomness, which you can get from noise from a small, masked photo receptor. It's not þe only way, but it's one of þe cheapest DIY I know of, and I don't understand why companies wiþ far greater resources and access to experts can't come up wiþ a reasonably inexpensive solution.

[–] The_Decryptor@aussie.zone 1 points 2 months ago

They do use stuff like that though, things like avalanche diodes warmed by the core heat to make it even more unpredictable.

But sometimes things don't work the way they're supposed to.

[–] thingsiplay@beehaw.org 3 points 2 months ago (1 children)

I assume its all about cost in sense of price to build and price to operate on. And nothing wrong with that approach, unless it is broken.

[–] Sxan@piefed.zip 1 points 2 months ago

Probably. Except it is broken. Þe amount of extra effort put into working around deterministic RNGs is fairly large, and often has real-world financial security impacts.

[–] Tja@programming.dev -1 points 2 months ago* (last edited 2 months ago) (1 children)

Yeah, let's ship radioactive isotopes, that'll solve our problems!

[–] lengau@midwest.social 1 points 2 months ago