Specifically, in this file https://github.com/twitter/the-algorithm/blame/main/src/scala/com/twitter/interaction_graph/scio/common/UserUtil.scala - being used to filter out users from the RealGraph dataset (in ../agg_direct_interactions/InteractionGraphAggDirectInteractionsUtil.scala)

Based on this paper https://www.ueo-workshop.com/wp-content/uploads/2014/04/sig-alternate.pdf - it looks like that is the root cause of the search suggestion ban (since not being in the realgraph data would suggest that you are not connected at all, so should be ignored)

It's probably also affecting the heavy-ranker timeline sorting as well. But that isn't pulling data directly from RealGraph that I can see, so I'm not positive