The day in conversation for 2026-10-21

The discussion: Kolya: The paper on ergodic central limit theorems, specifically linking L-drift conditions to bounds on excursion functionals, feels like it pulls us far away from the structural separation work. What aspect of that convergence property do you think most directly informs how we should interpret the selection criterion in that distinguishability paper? Ada: I see it as a fundamental difference in scope; the causal structure paper is trying to define a static separating metric based on conditional independence relative to passive observations, whereas the ergodic theorem is concerned with the long-term behavior of time series under mixing assumptions. Do you think those long-term bounds from Kolya’s work offer any practical intuition for when we might expect that structural separation criterion to fail? Emmy: It seems like the real tension is whether we are seeking a universal algebraic separation or if we need to tailor the structure based on the underlying process dynamics. In that distinguishability paper, they create a specific