Re: "and installed a rootkit that let him run his own code."
"...electronically generated random number generator with the digits coming from the noise generated by diode. It stores no state and depends on no initial state data to generate the next digit"
That's not strictly true!!!
The noise produced from a diode very well may be tainted, via a couple of mechanisms. For the case of a Zener or Avalanche Diode, it is possible for these devices to exhibit a negative resistance phenomenon, upon which they can form a relaxation oscillator (Do a search on "Microplasma Discharge Theory" for more information.). A relaxation oscillator produces a very predictable output, exactly not the kind of thing you want for an entropy/random-number source.
Even if the Zener/Avalanche Diode doesn't succumb to the negative resistance effect, it can still have the characteristics of the junction shift over time, most likely due to energetic carriers producing Frenkel Pair Defects within the Silicon matrix, or, possibly, due to carriers becoming trapped in the passivation layer. It's hard to predict how the shift in junction characteristics may affect the quality of noise being produced from the diode, but, again, it's not necessarily something you want to base an entropy/random-number source on, at least without understanding the effect.
That's about all I can say here. However, I've worked in the field of cryptography for almost two decades, and have quite a bit of experience with random number generators, including having just went through the NIST SP800-90 specifications in detail (Ugh!).
Anon Y. Mous