Sorry if this gets talked about a lot but is there common guidance for a dataset where we know experimentally that post-filtering is more performant but we also would like to apply pagination which is most easily done against our postgres data source rather than loading more data into memory than needed and paginating in-app? I think Materialize is one answer, but if it isn’t the only answer then I am curious what other guidance is out there. I think to some extent I could get away with grabbing more data than needed by a page in the initial sql query, filtering by SpiceDB check, and then further limiting the result to the page size requested from my API. Feels a bit clunky, but would be more performant than making iterative sql requests until enough data passes SpiceDB checks to fill a page.