Unfortunately, arthropods in general tend to be tough places for identifiers to start off, due to the sheer number of family/genera/species involved, the high number of superficial lookalikes (even across different families), the frequent need for details not captured in the average iNat observation, and the high frequency of observations with incredibly poor photo quality. The CV only further complicates the issue by introducing overly specific initial IDs that fledgling identifiers may not feel comfortable disputing.
Also, much of the publicly available identification information (if it even exists for a given part of the world or a given taxon - the US/Canada tends to be luckier than other parts of the world due to the existence of sites like BugGuide) tends to require having narrowed things down to the correct family or genus first, which is a separate skillset from IDing to species within a family or genus. Any helpful rules of thumb for higher-level sorting often have weird exceptions. The guides that do exist can be dauntingly long (for instance, this guide covers just identification of North American orb weaver species and only in the context of ventral-only views, and is quite lengthy, and doesn’t exhaustively cover all species).
Mentoring/training and/or creating new identification guides is also a time-intensive activity, and I can barely muster the motivation to go onto iNat as it is.