Five Bits Clean, Ten With Classical Help: Correcting the Headline on Our Own Quantum Result
A cryptographer with fifteen years in the field — eight of them as a U.S. Air Force cryptographer — read our quantum cryptanalysis preprint and wrote that our headline number was misleading.
He was right, and the correction is worth making in our own words rather than leaving it to a footnote in somebody else's article.
What we actually did
In July 2026 we ran a Simon-algorithm attack against the Even–Mansour cipher structure on IBM's Heron-generation quantum hardware. This followed an April 2026 result in which a team including Germany's federal security office broke the same cipher on real IBM hardware.
Our paper's title says 10. Here is what that means, precisely:
- N = 5 — clean quantum recovery. The machine hands you the key outright. The previous public record on real hardware sat at roughly N = 4. So the honest measure of pure quantum progress is one bit in three months.
- Up to N = 10 — hybrid. The quantum machine narrows the field to a shortlist of candidates, and a conventional computer picks the key out of that list.
Both are real results. They are not the same result, and a title that says "10" without saying which one invites the reader to assume the stronger claim. That is a fair criticism and we are adopting the distinction going forward.
Two further points Addie LaMarr raised in How Close Is a Quantum Computer, Really? (13 August 2026) are also correct: this is a preprint that has not been through peer review, and it comes from a commercial group rather than a national cryptographic agency. We state the pre-peer-review status in the paper itself. Treat the numbers as provisional. We do.
What this does not mean
It does not mean anyone is breaking encryption you use.
Even–Mansour at N = 10 is a practice cipher at a size no product deploys, and the attack assumes conditions — including a specific kind of query access — that fielded systems do not grant. The distance between this and RSA-2048 is not a matter of a few more bits.
LaMarr's own conclusion is the one we would write ourselves: the frontier is genuinely moving, and it is nowhere near anything you run. His article is a skeptical piece about how quantum progress gets overstated. We show up in it as a data point, not as a vindication, and that is the correct way to be cited.
Why post this
Because the alternative is worse.
If we let a headline number stand that a qualified outside reader has publicly called misleading, then every subsequent number we publish carries that discount. Cryptography is a field where the people evaluating your claim are, by disposition, looking for the place you overstated. The cheapest way to keep a result credible is to draw the line yourself, before someone else draws it for you.
So: five bits clean, ten with classical assistance, on real hardware, pre-peer-review. That is the claim. It is smaller than the headline and it is the one we can defend.
VIDRAFT Inc. is a deep-tech company founded in March 2024 and headquartered in Seoul, South Korea, developing Pre-AGI artificial intelligence models and quantum computing technology in-house.
