Source-linked quantum computing news and analysis on hardware, research, applications, post-quantum security, business and policy.
Publication date where available; source report or sources-checked date otherwise.
A 48-qubit experiment found that a local adiabatic circuit could beat an exact fermionic Fourier transform once noise entered the picture, exposing a blind spot in gate-count optimization.
A new proof shows that adaptive single-copy measurements can learn an exact stabilizer state with the same linear copy scaling as joint Bell measurements, without coherent access to two copies at once.
A new theorem gives constant expected error for private two-party distance estimation, but only when each participant returns the prescribed quantum message and completes the protocol.
A spiking neural network nearly matched a full-trace classifier on stored five-qubit readout data while producing updated answers as each 100-nanosecond chunk arrived.
A deterministic classical algorithm can estimate any chosen output probability for fixed-depth, bounded-fan-in circuits in polynomial time, even with arbitrary connectivity. Sampling remains open.
Concurrent proofs put static equilibrium properties in polynomial-time classical reach for a fixed-size impurity, while dynamics can still capture universal quantum computation.
Researchers detected distinct RF patterns from four generated quantum states outside an unshielded prototype. A closed rack cut the measured leakage by 21.2 dB on average.
A distributed lattice-surgery design reaches a 50% interface-erasure threshold only when local noise is absent. At one error in a thousand, the advantage largely disappears.
A superconducting-qubit study found that a tested adaptive filter reduced trace noise without resolving a fidelity gain. Choosing the measurement window did more.
A new theorem separates a quantum method from every polynomial-time classical algorithm in its oracle setting. The oracle is the point, not fine print.
A new beta lets selected teams explore error-aware programs. It is preparation for future dual-rail hardware, not access to that hardware today.
Six projects move into demonstrators. Energy forecasting and a grid clock show why practical quantum progress needs more than a funding headline.
A new superconducting experiment turns selective photon loss into a useful ingredient for cat qubits. The protected qubit itself is still a next step.
A new memory architecture separates storing quantum information from manipulating it. The distinction may matter as much as adding another qubit.
A new survey reports concentrated maintenance in many sampled quantum-software repositories. It raises a human infrastructure question, not a verdict on every project.
A Singapore comparison of lutetium-ion clocks makes small differences in height part of the measurement. The story is about gravity as well as time.
A small photonic experiment survived a trip to orbit. Its struggle with sunlight may be more revealing than the launch.
A quantum experiment combines light signals and strict timing to test where a device actually is. It is not a replacement for GPS.
A gallium-arsenide experiment reveals how a crowd of electrons can carry charge in unexpected portions. It is not a photograph of an electron being cut up.
A cold crystal preserved quantum correlations while information waited in a spin state. That is a useful component for a future network, with more work still to do.
A wireframe SQUID on a cantilever tackles the geometry of measuring very small magnetic signals.
Classical shadows help researchers estimate useful properties without reconstructing an entire quantum state.
A quantum model can have adjustable settings and still struggle to find a useful way to improve.
A federal award of up to $1 billion puts quantum manufacturing, not just qubit design, in the spotlight.
Page 13 of 19
A little perspective on a fast-moving field.
