I’d say you have to know your audience but it seems to work fine here. Plenty of self hosted blog posts make it to the front page of HN. I don’t see anyone writing for the broader masses and maybe that’s more the type of writing you think will be quashed, but writing for a niche seems to proliferate to the specialty sites just fine.

" writing for a niche seems to proliferate to the specialty sites just fine."

Does it, though? I'd actually like to see a thorough study done on this. My hypothesis is different from yours: the proliferation of "specialty sites" is a product and symptom of the walled garden phenomenon, and further serves to sequester and censor undesirable content from the wider audience.

I'm pretty happy that i havent had to read celebrity news sites because i have an algorthmically light diet, especially in the last few weeks with Will Smith etc. Do you really think this nichemaking is bad inherently bad? The natural course is that good niches fill and grow until they become the mainstream and the folks who are nudged out go make new niches. this even happens with Big Media... It even happened here.

What are you suggesting is better, though? really, that we all just publish one of 10000 articles to dev.to and hope their search and popularity algorithm puts your article in front of peoples' eyes? That everyone should upload video to a single time-ordered feed that you have to scroll through until you find something you like without algorithmic involvement? Im seriously confused, how do you do content selection or personal curation in a world without niches

> Many people barely see a browser window, and consume all online information inside the walled gardens.

I seem to recall a survey in South East Asia asking people about their internet use, and there were more Facebook users than internet users...

Ah, here: https://qz.com/333313/milliions-of-facebook-users-have-no-id...

"11% of Indonesians who said they used Facebook also said they did not use the internet."

Such survey results cannot be used to infer anything without understanding how the survey was worded. For instance, if one of the answers were "I use Facebook, but not the internet" then A) some people will choose it just by virtue of it being there, and B) it could occur to some people that the survey was poorly worded but meant "I do not use the internet for anything other than Facebook".

I remember a rash of such loaded survey questions going around my country a few years ago, it looked like it was specifically designed to be answered with answers that could be used to slander my nation with "N% of these people think this" results. It's usually not disinterested parties investing the time and resources to run these survey, rather, it's usually people for whom the investment will pay off if they can get the results they want.