Classical simulation closes in on another advantage claim
A tensor-network method reproduces sampling results that were argued to be out of classical reach.
By Ines Duarte
Written in-house by the Analysis desk — an explainer, not a report of a news event.
- Simulation matched the published fidelity at lower cost than expected.
- Advantage claims increasingly hinge on runtime accounting.
- Error-corrected workloads remain the durable target.
Every advantage claim eventually meets a better classical algorithm. This time a tensor-network contraction scheme reproduced the sampling distribution at a computational cost far below what the original analysis assumed.
The pattern is by now familiar and arguably healthy: sampling claims are fragile, and the durable target has moved to error-corrected workloads where the resource accounting is harder to undercut.
Members in our Discord ran the numbers on the reported cluster hours and the thread is worth reading for the accounting alone.
This story started as a thread in our Discord. Join the discussion with the researchers and engineers who work on it.
Related
Entanglement distributed across a live metropolitan fibre network
Not a dedicated dark-fibre link — the demonstration ran alongside ordinary classical traffic on deployed infrastructure.
Surface-code processor stays below threshold as it scales to 1,000 qubits
The logical error rate keeps falling as the code distance grows — the clearest sign yet that adding qubits now buys reliability instead of noise.
Open-source compiler cuts logical-qubit overhead by a third
A new lattice-surgery scheduler squeezes routing overhead out of fault-tolerant circuits, and it ships under a permissive license.