Migrating to Post-Quantum Crypto? You Still Need Real Randomness.

1. What quantum computers actually break, and don't

Shor's algorithm breaks the math RSA and elliptic-curve cryptography depend on: factoring large integers and solving discrete logarithms. That's the entire reason post-quantum migration exists: those problems stop being hard once a large enough quantum computer runs Shor's algorithm against them. Grover's algorithm affects symmetric crypto and hashing, roughly halving their effective security, which is why post-quantum guidance calls for bigger keys (AES-256 over AES-128, SHA-384/512 over SHA-256).

Neither algorithm has anything to do with how a random number generator produces unpredictability in the first place. A weak RNG is a weak RNG whether the attacker has a quantum computer or not. A strong physical entropy source stays strong regardless of what kind of computer eventually tries to break the cryptography built on top of it.

2. Post-quantum algorithms need more randomness, not less

Lattice-based schemes, ML-KEM (Kyber) for key exchange and ML-DSA (Dilithium) for signatures, sample from probability distributions over large polynomial rings using rejection sampling and discrete Gaussian sampling, consuming substantially more random bytes per key or signature than RSA or ECDSA ever needed. Hash-based signatures like SLH-DSA (SPHINCS+) generate huge numbers of one-time key pairs, bottlenecked entirely on a genuinely fresh, unpredictable byte stream. A biased or predictable RNG under these schemes isn't a theoretical concern: it's the same class of failure that made ECDSA nonce reuse catastrophic, now against algorithms with more randomness to get wrong.

3. Wait, isn't RF and cosmic noise quantum too?

At the deepest level, yes. Radio waves are photons, and thermal (Johnson-Nyquist) noise ultimately comes from the quantum-mechanical motion of electrons. Nearly everything physical is "quantum" once you go down far enough. That's not actually the distinction that matters here.

Dedicated quantum RNG hardware isolates a single quantum measurement: a photon hitting one side of a beam splitter, or a vacuum-fluctuation reading, whose outcome is random by the Born rule itself, a formal guarantee from quantum mechanics independent of how complex the surrounding system is. Galactropy instead harvests unpredictability from a macroscopic, high-dimensional thermal and RF environment: the combined chaotic motion of billions of electrons and the ambient radio spectrum, sampled by ordinary SDR hardware. Both are physically real and both are unpredictable in practice; only the former comes with a specific, certified quantum-randomness guarantee. That's the actual distinction worth knowing about your entropy source, not whether quantum mechanics is "involved" somewhere in the physics, which it always is.

4. Verifiable today, still verifiable after Q-day

Galactropy's audit trail is currently signed with Ed25519, which is itself not post-quantum secure, like most of the signatures securing the internet today. When post-quantum signature migration reaches infrastructure like this, the audit chain moves with it. What doesn't change is the source: physical radio noise stays unpredictable regardless of how far quantum computing progresses. That's a different property from "the signature on this particular record happens to be classical today," and it's the one that actually matters for the entropy itself.

5. Where this actually helps

A few concrete situations where mixing in an independent, verifiable entropy source earns its keep during a post-quantum transition:

A minimal example: pull entropy and mix it into the pool before generating keys:

curl -s "https://galactropy.com/raw?bytes=4096" >> /dev/urandom

Then run your post-quantum key generation as usual. liboqs, OpenSSL's oqs-provider, or whatever library you use will draw from the now-augmented pool automatically.