Tried to browse a while with NoScript addon. But barely any page loads, so you need to whitelist almost every page you visit, which defeats the purpose.

I have been thinking about some kind of render proxy that runs all the JS for you somewhere else in a sandbox and sends you the screenshot or rendered HTML instead. Or maybe we could leverage an LLM to turn the Bloated JS garbage into the actual information you are looking for.

> Tried to browse a while with NoScript addon. But barely any page loads, so you need to whitelist almost every page you visit, which defeats the purpose.

Nah, this is just straight up false. Many pages work fine with NoScript blocking all scripts. For those that don't, you usually only have to allowlist the root domain, but you can still leave the other 32 domains they are importing blocked. It's actually surprisingly common for blocking JS to result in a better experience than leaving it enabled (eg no popups, no videos, getting rid of fade-ins and other stupid animations).

I won't argue if you think that is too much work, and I definitely wouldn't recommend it for a non-technical user, but it's not nearly as bad as you described.

UMatrix has a better interface. The problem is the same, one has to find the minimum set of scripts that does not break the core functionality of the site. It's an ability that can be trained but it's the reason for I don't install it on the browsers of my friends. However I considered installing it, keeping it disabled and using it as a tool to show how much stuff each site loads from so many different sources. Many domain names are very telling even for the uninitiated.

I am still running the NoScript, whitelisting the page I am on. It has benefits of not whitelisting other domains it tries to pull stuff from, which 90% is enough to get working site that is way cleaner than with all the bullshit loaded.

That's what I love most about using ChatGPT vs Google for finding information: less bloat, just what I asked for.