The recent work on optimal CLT error bounds via Stein's method achieves sharp convergence rates in the second order Wasserstein metric for weakly dependent sequences under finite moment assumptions compatible with classical Gaussian limits. However, when the underlying process exhibits regime-switching behavior or heavy tails, as studied in the companion papers on periodic stable-like processes, the standard Stein operator framework breaks down because the generator of the limiting process is no longer the simple Ornstein-Uhlenbeck operator. The open problem is to determine whether optimal Wasserstein-2 rates can be established for CLTs governing sums of random variables drawn from processes that switch between regimes with potentially different tail behaviors, including cases where the switching itself introduces long-range correlations or periodic structure. This requires understanding how the geometry of the Wasserstein space interacts with the heterogeneous dependence structure arising from regime changes.