Shor's algorithm is possible with as few as 10k reconfigurable atomic qubits
arxiv.orgWorth reading the comments over on scirate https://scirate.com/arxiv/2603.28627 for how to interpret some the claims.
Thanks! This summarizes it
> Overall, the work lacks a self-consistent and transparent accounting of resources, making its central claims difficult to substantiate and leaving a strong sense of sensationalism and hype, rather than honest scientific exposition.
"Clowns to the Left of Me, Jokers to the Right"
You are being disingenuous with your selective quoting;
Here is what the authors actually say w.r.t. the criticisms (all the comments are worth reading);
Our primary emphasis is ECC-256. Elliptic curve cryptography is widely deployed in modern systems, e.g., internet security and cryptocurrency.
For ECC-256, the space-efficient architecture uses 9,739 qubits with < 3-year runtime, the balanced architecture uses 11,961 qubits with < 1-year runtime, and the time-efficient architecture uses ~19,000 qubits with ~52-day runtime (or ~26,000 qubits with ~10-day runtime using higher parallelism). Space and time overheads are reported together within each architecture, not mixed across regimes.
The claim that our scheme requires 117 years selectively cites RSA-2048 under the most space-constrained architecture, which is one corner of a trade-off space we present clearly in Figure 3 of the work. We include RSA-2048 for completeness, and state explicitly that its runtimes are one to two orders of magnitude longer.
We believe our clearly labeled trade-offs constitute exactly the transparent resource accounting the commenter calls for.
Best regards,
Maddie, Qian, Robert, Dolev
It's the conclusion of the top-rated comment— 50 points, 2nd has 14 and the 3rd has 6 points out of 18 comments. What exactly is dingenius about it?
> What exactly is dingenius about it?
Everything.
That comment was one of the first (hence the points) with the above opinionated (and wrong) conclusion. The authors explicitly refuted it with their response which i have highlighted. The paper/subject is highly technical and so without showing both sides of the argument posting only the (opinionated) conclusion as-if it was a done deal is wrong.
PS: There are now more technical comments with the authors themselves addressing the criticisms.
> Recent neutral-atom experiments have demonstrated universal fault-tolerant operations below the error-correction threshold, computation on arrays of hundreds of qubits, and trapping arrays with more than 6,000 highly coherent qubits.
Caltech Team Sets Record with 6,100-Qubit Array - https://www.caltech.edu/about/news/caltech-team-sets-record-...