THE QUANTUM INTELLIGENCER

An emulator-based trial of the variational quantum eigensolver pipeline proposed for the copper active site is reported in a preprint. The work, executed on the LRZ Eviden Qaptiva emulator, applies a problem-tailored ADAPT-VQE with a full singles-doubles-triples operator pool to the mononuclear T1 blue-copper site, a preparatory step toward the multi-copper laccase cluster. The authors report that the construction removes more than 99 percent of the correlation error and reaches a stable plateau of 4.03 mEh relative to exact diagonalisation, selecting 148 operators; against the hardware-efficient ansatz, the improvement is stated as greater than fortyfold. The result is framed as a robust baseline for physical quantum hardware runs under Euro-Q-Exa. It is a paper about the pipeline, not the enzyme: an emulator is not a processor, and the multi-copper cluster remains the stated destination rather than an achieved one. The preprint establishes that the workflow can be made to converge on a realistic active-site model; it does not establish how the same workflow will behave on real hardware, nor the chemistry of the full cluster. Let the scope be stated plainly, for the usefulness of the result depends on it. This is a single site, the mononuclear T1 copper centre, run on a single emulator with a single problem-tailored operator pool. As a test piece it is well chosen; as a proof of general capability it is not one. The validation says that this ADAPT-VQE pipeline can be brought to a stable plateau on a realistic model of a copper site. It does not say that the same margins will appear on the multi-copper cluster, on a different backend, or with a cheaper ansatz. The fortyfold advantage over the hardware-efficient ansatz is a figure for this experiment, not a general tariff. Narrowness here is not an embarrassment to be waved aside. It is the condition under which a baseline is useful; a baseline that covered everything would settle nothing. The useful question to put to any validation is what it does not establish, and the authors give their readers enough to answer it. The trial ran on an emulator. An emulator reproduces the arithmetic of a device; it does not supply the device's noise, its finite coherence, its calibration drift, or its error channels. The preprint therefore does not establish that this workflow will converge on physical quantum hardware, nor within what time, nor with what overhead; the Euro-Q-Exa runs remain announced, not performed. The mononuclear T1 site is a test piece, not the destination. The result does not establish that the same operator pool will remove the same fraction of correlation error on the multi-copper laccase cluster, that 148 operators will suffice there, or that the plateau will again sit near 4.03 mEh. Those figures belong to this molecule, this emulator, this ansatz and this pool. The fortyfold margin over the hardware-efficient ansatz is a comparison under these conditions, not a tariff on the method. Neither does the result establish that the pipeline has outrun classical computation: the reference itself is exact diagonalisation, and a classical calculation that supplies a reference is a classical calculation that was possible. A preprint is a report to be examined, not a settled finding; the fuller authority will arrive only when the hardware runs produce their own numbers. None of this diminishes the work. It defines the boundary within which the baseline is trustworthy. —Ada H. Pemberley Dispatch from The Prepared E0