I observed this kind of behavior in many papers nowadays. This extremely painful for research, because some better candidates could be overseen and FAANG publishs a majority in the ML-paper section. Its a mess.