I think that the tradeoff in terms of getting feedback and providing something for others to build on, especially others outside of the narrow EA-motivated community, is often worthwhile.
This should be obvious for everyone! As an outside observer and huge sympathizer, it is super-frustrating how siloed the broad EA/rational/AI-alignment/adjacent community is --this specific issue with publication is only one of the consequences. Many of "you people" only interacting between "yourselves" (and I'm not referring to you, Davids), very often even socially. I mean, you guys are trying to do the most good possible, so help others use and leverage on your work! And don't waste time reinventing what is already common or, at least, what already exists outside. More mixing would also help prevent Leverage-style failures and probably improve what from the outside seems like a very weird and unhealthy "bay area social dynamics" (as put by Kaj here).