If you're trusting some closed source black box silicon chip printed in closed unauditable uninspectable fabs by a bunch of closed GovCorp entities, all of which have ongoing history ranging from broken to malicious products... to "generate" and feed you "random numbers", and further trusting the box to tell you whether or not itself is working as "advertised", and that the numbers it's feeding you were are and will be "random", etc... then yes, you're doing it very wrong. People can add it as one of many sources into whatever their favorite [un]trusted opensource hash chain PRNG software is, the adding simply serving as redundancy in case failure of N-1 sources. Yet regardless, if none of those sources into the XOR's are TRNG's, and if there was no trusted sample file of random that had already passed all desired hardness checks included in the addition as bootstrap against all-source and startup settling failure... it's a bit moot. Most major open OS... BSD Linux, and presumably closed Win and Mac, already fixed those basics not long ago. However it remains rightly and only your responsibility to do the last remaining thing... All over the internet there are reference designs for opensource True RNG's using discrete testable logic components in the output stages that anyone can build and plug into their serial port as one of the sources, so go do that. Use Shannon, Quantum, Private Observations of Nature's Ephemerals, Logic, Continuous Lifetime Testing, etc as your spirit guide. Or at least keep mashing keyboard and mouse. Then you have seed mix, and at least up to whatever bit-equivalent strength your favorite [un]trustable opensource PRNG is, including possibly for XOR onetime pad. This sort of RNG theory has been known for many many decades. It's utterly ridiculous 20+ years since the first opensource OS/library RNG exploit papers, that whack-a-mole and accepting the "oopsie" excuse are still the case. And futher ridiculous that no one has launched into the profitable world-changing greenfield that is... #OpenHW printed in #OpenFabs under #OpenAudit https://en.wikipedia.org/wiki/Random_number_generation https://en.wikipedia.org/wiki/Hardware_random_number_generator https://en.wikipedia.org/wiki/Cryptographically-secure_pseudorandom_number_g... https://en.wikipedia.org/wiki/List_of_random_number_generators https://en.wikipedia.org/wiki/One-time_pad https://en.wikipedia.org/wiki/Provable_security https://en.wikipedia.org/wiki/Information_theoretic_security https://en.wikipedia.org/wiki/Information_theory https://en.wikipedia.org/wiki/Claude_Shannon https://en.wikipedia.org/wiki/Post-quantum_cryptography