Replica symmetry breaking in multimode-cavity spin glasses
In plain words
Atoms in a cavity with many light modes interact with random signs, like the frustrated magnets called spin glasses and like associative-memory networks. The question is whether they show the hallmark of a true spin glass.
Precise statement
Transversely pumped atoms in a confocal multimode cavity realizing all-to-all couplings $J_{ij}$ with random signs (Sherrington-Kirkpatrick or Hopfield type). Measure the overlap distribution $P(q)$ between replicas (repeated preparations with the same $J_{ij}$) and determine whether it has continuous support (replica symmetry breaking) and whether it reflects equilibrium or a driven-dissipative steady state. The equilibrium Parisi solution is the comparison, not the question. Answer: $P(q)$ with an independent test of equilibration.
What would settle it
Replica-overlap measurements across many disorder realizations with checks of fluctuation-dissipation relations.
Status in the literature
Unverified note
Replica symmetry breaking and an ultrametric overlap structure were observed in a driven-dissipative vector spin glass (Kroeze et al., Science 2025), and an Ising cavity spin glass was realized (Marsh et al., PRL 2025); the relation to equilibrium Parisi theory is open.