That's the fundamental underlying reason behind all the rules and all the deletion discussions. It's not that the article isn't important, it's not that the content is wrong or not useful or that we don't want it. It's that Wikipedia would be pure chaos if it let people just write without sources, and the fights over content would never end.
Notability just boils down to "can we find enough reliable sources to write an article we can verify and stand behind?". Just quoting what authors of the topic say about their own work would of course fall a little short of what people expect from an encyclopedia.
what's the gate being kept and by whom?
This is what makes them reliable in Wikipedians' judgement. A publication that regularly made errors, or published claims regardless of veracity, or left errors in place even after being notified of them, would quickly lose its "reliable" badge. Wikipedians even track the downfall of a reliable source becoming an unreliable one, and say you can only cite articles from the time period the source was reliable.
And Wikipedia itself is ready and waiting to remove citogenesis and prevent it coming back, the moment it's no longer in a reliable source. So while citogenesis can happen, it's detected and resolved over time, it's not some "gotcha".
Yes, for the encyclopedia project. But they are many use cases where saving who said what, when, within which contextual summarizable frame can also prove very useful.
Luckily we still have things like archive.org still covering a large part of such a need. Because otherwise, anyone can say anything and its contrary just before denying reality never happened, and then the web becomes saturated of every assertion and its contrary being all covered in some farmed nonsense without any way to weight them in the hand of general public.