But it's kind of a vicious cycle: users avoid bad sites, traffic drops, sites shove in more ads to survive, UX gets worse, and so on

> sites shove in more ads to survive

This is where it breaks down; why would they shove in MORE ads when their readers are going down? I'm not saying it's a rational decision, of course.

I suspect a big part is metrics-driven development; add an aggressive newsletter popup and newsletter subscriptions increase, therefore it's effective and can stay. Add bigger / flashier ads and ad revenue increases, therefore the big and flashy ads can stay.

User enjoyment is a lot harder to measure. You can look at metrics like page visits and session length, but that's still just metrics. Asking the users themselves has two problems, one is lack of engagement (unless you are a big community already, HN doing a survey would get plenty of feedback), two is that the people don't actually know how they feel about a website or what they want (they want faster horses). Like, I don't think anybody asked Google for an AI summary of what they think you're searching for, but they did, and it made people stay on Google instead of go to the site.

Whether that's good for Google in the long run remains to be seen of course, back when Google first rolled out their ad problem it... really didn't matter to them, because their ads were on a lot of webpages. Google's targets ended up becoming "keep the users on the internet, make them browse more and faster", and for a while that pushed innovation too; V8, Chrome, Google DNS, Gears, SPDY/HTTP/2/3, Lighthouse, mod_pagespeed, Google Closure Compiler, etc etc etc - all invented to make the web faster, because faster web = more pageviews = more ad impressions = more revenue.

Of course, part of that benefited others; Facebook for example created their own ecosystem, the internet within the internet. But anyway.

Unless there's a conscious reset, like the Onion reboot. Now with physical copies!

Doesn't scale, but maybe that's the only way to survive.