This dovetails into an idea I had [0]. Basically just client side scrape the web as it's used and deliver people this plain text and simple forms. It would have a maintained set of definitions and potentially even logic to put a better "front" on all this bullshit. It's like reverse ad block where you only whitelist some content instead of blacklisting it. You could argue sites will get good at fighting it, but if used enough by the common user, they'd just alienate them (e.g. my scrape/front for Google search makes it clear which results the app has a friendly scrape/front for).