Does the pseudo log-likelihood from PROC GLIMMIX mean anything?

[SAS_Cares]

A Twitter user asks:

SAS users: does the pseudo log-likelihood from PROC GLIMMIX mean anything? I am specifying a marginal (R-side) model thus can't have AIC/true log-likelihood for model comparison. #statstwitter #epitwitter

— Andreas Teferra (@andreteferra) January 11, 2020

With some clarifying questions and answers in the thread.