This is a wild misrepresentation of the opposing perspective. Nobody is arguing that we can't discuss election fraud.
The argument—which I'm sure you are actually aware—is that there needs to be some level of credibility to the idea that a) fraud occurred, and b) that it happened in meaningful quantities before we spend significant time, cost, and effort in investigating claims.
Simply having lost is not a credible claim to investigate widespread fraud. Finding one or two isolated cases in elections with margins of thousands or more votes is not a credible claim to investigate widespread fraud.
Further, fraud cannot simply be a claim that is made and then perpetually reinvestigated by decreasingly-reputable third parties until you are able to invalidate an election whose outcome you disagree with.