What “No Quantum Advantage” Means for a 6-Qubit Toy Network
The Resilience Lab states clearly that no quantum advantage is claimed. For a six-node toy network with six decision qubits, that statement is both accurate and necessary. This article explains what the phrase means in this specific setting and what a reader should not infer from a successful run.
Advantage is a comparison against the best known classical method on a relevant problem
Quantum advantage, in the sense used in research literature, is a demonstration that a quantum device or algorithm solves a problem of practical interest faster or with better quality than the best classical algorithm known for that problem, under stated resource bounds. It is not the observation that a quantum circuit produced a feasible answer. Many classical methods also produce feasible answers.
On the Resilience Lab instance, the classical exhaustive search examines all 15 feasible selections and returns the optimum in well under a millisecond. A QAOA simulation that samples a few hundred shots and finds a matching or near-matching score has not shown an advantage. It has shown that the circuit and the sampling procedure can locate a good solution on a problem that is already trivial for classical enumeration.
Why the lab is sized this way
Six qubits produce 64 amplitudes. That size is small enough to simulate exactly in the browser with a statevector, to test every energy value against an independent calculation, and to keep the classical baseline instantaneous. It is also small enough that any claim of advantage would be false. The lab is built as a research preview for understanding QAOA mechanics, constraint handling, and reproducible comparison, not as a benchmark for supremacy or practical speedup.
What a good run still demonstrates
Even without an advantage claim, a run can show several useful things: that the phase-separation and mixer gates are applied correctly, that the classical angle optimizer improves the expected energy, that Born-rule sampling produces a distribution consistent with the final amplitudes, and that the penalty term discourages but does not eliminate infeasible bitstrings. Those are properties of the implementation and of the algorithm on this instance. They are worth verifying. They are not evidence that the same approach will be faster than classical methods when the number of candidate links grows into the thousands.
How the page keeps the boundary visible
The lab places the classical optimum next to the best QAOA sample, labels the problem as a toy network, and states that the simulation is noiseless and local. The downloadable evidence includes the seed, depth, shot count, and both results. A reader who wants to explore larger instances will need a different classical method and a different scale; the current page does not provide that scale and does not claim to.
The absence of a quantum-advantage claim is not a limitation of the write-up. It is a direct consequence of the problem size the lab was designed to explore.