Quantum computing papers, judged / Independent · source-traced
Most quantum announcements are noise. This is the signal.
One scientific paper at a time: what it shows, whom it concerns, and what is still missing before anyone can use it — in plain language, with the quotation it rests on. No physics degree required.
- EvidenceA verbatim quotation from the paper in every brief
- VerdictReal today · Milestone · Plausible by · Beyond current roadmaps
- Length600 to 800 words. No filler.
- PrerequisitesNone. Terms explain themselves where they appear.
Latest briefs
The newest papers judged, across all fields.
-
Quantum ML: 26 of 29 top papers aren’t ready
We read 29 of the most-cited quantum machine learning papers. 26 show progress but state no measurable distance to a real business task.
-
Lightweight cipher fails NIST’s quantum test
A paper calculation finds a small-device cipher falls short of NIST’s quantum security bar. It shows a design gap, not a break of any deployed system.
-
Quantum attacks on Chaskey are far off
Breaking Chaskey or PRINCE takes 2⁶⁵ operations, against 2³¹ for RSA-2048. Systems with long-term secrets should plan migration now.
-
Cracking RSA with fewer qubits takes months
Factoring RSA-2048 would take 177 days on 13,436 qubits plus a memory not yet built. Migration deadlines do not depend on it.
-
Harvard and QuEra run 48 protected qubits
A 280-atom machine scored up to an order of magnitude above unprotected qubits on test circuits. Commercial jobs need far more qubits.
Chemistry, drugs and materials
Simulating molecules and materials: what a quantum computer would need, and how far that is.
-
A drug enzyme needs 4.6 million qubits
Modelling one drug-metabolism enzyme takes 4.6 million qubits and 73 hours, on paper. No announced roadmap reaches that; use it as a planning yardstick.
-
No quantum roadmap reaches battery design
Simulating one battery electrolyte molecule needs 16,382 error-protected units; no announced machine reaches that. Battery makers stay on classical tools.
-
One molecule needs four million qubits
A paper with Google Quantum AI sizes one FeMoCo simulation at four million qubits and 3 days. No announced roadmap reaches that; nothing to buy today.
Optimization
Finance, logistics, scheduling: where a quantum method is claimed to help, and what the papers actually show.
-
Goldman Sachs cuts quantum pricing cost 16x
Goldman Sachs authors cut the work 16-fold, yet pricing still needs 4.7 k logical qubits at 45 MHz. Banks can’t use it yet; no such machine…
Quantum machine learning
Machine learning on quantum computers, and machine learning used to build them.
-
Quantum ML: 26 of 29 top papers aren’t ready
We read 29 of the most-cited quantum machine learning papers. 26 show progress but state no measurable distance to a real business task.
Error correction and logical qubits
The machines themselves: error correction, logical qubits, and the results that later estimates are built on.
-
Harvard and QuEra run 48 protected qubits
A 280-atom machine scored up to an order of magnitude above unprotected qubits on test circuits. Commercial jobs need far more qubits.
-
Google adds qubits and cuts errors by 4%
Google’s 49-qubit code erred 2.914% per cycle versus 3.028% for 17-qubit codes. Scaling helps, but hardware must still improve by at least 20%.
-
QuEra’s atom bet reaches 99.5% accuracy
Harvard and QuEra hit 99.5% accuracy on up to 60 atoms at once, past the 99% error correction needs. A planning signal, not a usable machine…
Post-quantum cryptography
Attacks on today’s encryption and the methods meant to replace it.
-
Lightweight cipher fails NIST’s quantum test
A paper calculation finds a small-device cipher falls short of NIST’s quantum security bar. It shows a design gap, not a break of any deployed system.
-
Quantum attacks on Chaskey are far off
Breaking Chaskey or PRINCE takes 2⁶⁵ operations, against 2³¹ for RSA-2048. Systems with long-term secrets should plan migration now.
-
Cracking RSA with fewer qubits takes months
Factoring RSA-2048 would take 177 days on 13,436 qubits plus a memory not yet built. Migration deadlines do not depend on it.
The signal, by email
Get the next one.
One email when a new brief goes live — a verdict, the evidence behind it, and the sources. No digest, no sponsors, no third-party sharing.
Your address is stored for this newsletter only. Unsubscribe in one click. See the privacy page.