Making true random numbers with radioactive decay (2021) | Hacker News Reader

Making true random numbers with radioactive decay (2021)

partofthething.com

18 pointstheletterf6 comments

I love it! Stuff like reminds me of the wall of lava lamps that Silicon Graphics used to generate random numbers.

https://en.wikipedia.org/wiki/Lavarand

In my first job out of uni, one of the questions the interviewer asked me was how I would design an impromptu RNG system. It caught me by surprise, so I answered that I would stare unblinking for as long as possible, alternatingly chanting "0" and "1" until I blinked. It was an absurd answer, but I was a young kid, and it was the first thing I thought of. I got the job, but I doubt it contributed substantially.

To this day, if I sense that a prospective applicant is someone who seems to embody an innate sense of curiosity about the world around them, I'll toss this one out as a fun thought exercise.

"If the length of the first gap is bigger than the length of the second, you emit a 1. Otherwise, you emit a 0."

Doesn't radioactive decay slow over time? That would mean, in the (possibly very) long run, 0 would be slightly more likely than 1.

From the original https://www.fourmilab.ch/hotbits/how3.html

> To create each random bit, we wait until the first count occurs, then measure the time, T1, until the next. We then wait for a second pair of pulses and measure the interval T2 between them, yielding a pair of durations. If they're the same, we throw away the measurement and try again. Otherwise if T1 is less than T2 we emit a zero bit; if T1 is greater than T2, a one bit. In practice, to avoid any residual bias resulting from non-random systematic errors in the apparatus or measuring process consistently favouring one state, the sense of the comparison between T1 and T2 is reversed for consecutive bits.

That would do it. It's interesting that they compare 1-2 with 3-4 and not 1-2 with 2-3. I wonder if there's a reason for that.

If the time slices are 1-2 and 2-3 (or 3-2), will the timing of event 2 introduce a bias?

Currently its:

    1-2 vs 3-4
    7-8 vs 5-6
    ...

If event 2 is being used as common, would it then be:

    1-2 vs 2-3
    4-5 vs 3-4
    ...

And so with this second set, you can see that 2 is common to both times of the first bit and 3 is common to the first and second bit (and 4 is common to both times of the second bit).

If there is a systemic error in the system, having those be common timing events might make it worse? And I'm guessing there.

@vunderba, love the lavarand post - take a look at this - www.realrandom.co

Reply on news.ycombinator.com