The fact that the volume is exploding but the graveyard is also exploding, is a sign that the system is working, not that it's broken (the filter is working).

I did 3 ShowHN in 2024 (outside of the scope of this analysis), one with 306 points, another with 126 points and the third with... 2. There's always been some kind of unpredictability in ShowHN.

But I think the number one criteria for visibility is intelligibility: the project has to be easy to understand immediately, and if possible, easy to install/verify. IMHO, none of the three projects that the author complains didn't get through the noise qualify on this criteria. #2 and #3 are super elaborate (and overly specific); #1 is the easiest to understand (Neohabit) but the home page is heavy in examples that go in all directions, and the github has a million graphics that seem quite complex.

Simplify and thou shall be heard.

The difference since Clawd and friends became popular is palpable, you see it growing on GitHub too with the PR spam.

I'm wondering how much of it is portfolio building to keep or find a new job in a post-Ai coding world