C.29.3:11 - SoTA-Echoing
The selected answer combines a physical-realization comparison with scope-sensitive result use and a return to formulation when the available means suggest a different construction. The examples are constructed under their stated models.
| Working question | Source contribution and selected use | Comparison and limit |
|---|---|---|
| How does an abstract result become obtainable through a physical arrangement? | Adopt preparation, evolution and interpretation from Horsman et al. (2014), as used in :4.2–4.5. | Compared with abstract refinement alone, the comparison exposes input and readout failures. The theory’s general classification claims are outside this method’s required conclusion. |
| How does computation participate in physical control? | Adapt the compute-cycle/control-cycle distinction from Horsman et al. (2026) in :4.1 and :4.3. | The preprint sharpens the timing and output question. Its broader claim about all control systems is not needed to establish these worked cases. |
| Should the device or the computational model change? | Adapt Stepney’s 2019 co-design proposal and Kalita et al.’s 2026 substrate-to-model construction as the two-way return in :4.6. | A fixed-model implementation remains preferable when it meets the use at lower cost. The bosonic example demonstrates a particular methodology; it establishes no general performance advantage for every substrate. |
| How is a stochastic program realized? | Adapt the distinction between a target stochastic program, its compiled kernels and its interpreted readout from Amico et al. (2026), III and V, in :4.1 and :4.5. | Comparing local operations, composed output laws and receiving-task results can reveal different failures. The preprint’s compilation demonstrations do not establish a general hardware energy advantage. |
| Must physical evolution reach equilibrium? | Adapt the finite-time approximation choice from Thermodynamic natural gradient descent (2026), Results, in :4.3. | A useful inner estimate can support the learning update before equilibrium. The reported thermodynamic timing is estimated; select the interval by the receiving computation’s requirements. |
| How much additional checking is useful? | Use C.11.DUA to compare what a further observation, proof or trial could change. | A conditional design can be sufficient. An actual-performance or wider-range claim needs the grounds that its receiving use consumes. |
Reconsider the realization when a required input, observation relation, execution condition or receiving tolerance changes. A new substrate or computational model can also change which preparation, operations and interpretation are worth developing.