If people are assigned a "gatekeeping" position, they should at least be expected to be able to tell great minds from crack pots. (As opposed to just blindly applying a statistical filter)

Otherwise why not just replace them with a script (or these days, an AI)

Isn’t history replete with examples of great minds who were received as crackpots until, after careers’ worth of labor to validate them, some of their crazy ideas turned out to be correct?

The problem is the actual crackpots outnumber hidden gems by many orders of magnitude, so wasting time trying to sift through the noise is a losing proposition.

And I don’t know many, if any, examples of people received as crackpots that later were considered genius. At worst genius takes a little time to be recognized, which is a situation far more common.

I think maybe Srinivasa Ramanujan? Maybe not a full crackpot, but defo not worth spending time on because his work looked like Aliens came by and dropped a textbook.

I forget the quote, but one of the guys that recognized that he was worth looking into said something to the effect of "his work was too crazy for him to be crazy".

No one ever thought Ramanujuan a crackpot. Hardy immediately realized his genius and brought him to the UK.

In this context "gatekeeper" is often defined as broadly as "any individual academic or researcher receiving a cold email".

E.g., economics.

I've literally never met a person who didn't consider themselves economically knowledgeable.

It's like an analog to the fact that almost all people consider themselves funny.

Nothing is more fairly distributed than common sense: no one thinks he needs more of it than he already has. -- Descartes