I normally avoid posting on threads like this, because there's more pile-on then there is coherent discussion, but I think there are some things missed here, and I want to elevate the discussion.<p>Let's take a step back for a moment from the fact that this is extremely one-sided, since there are many sites used all over the web which have a view on most of your browsing traffic, like Google Analytics (try and pry that from the cold, dead hands of administrators), to reddit, to other ad networks. I'll assume that the reader would just respond that we should do this for all sites.<p>Think a bit about the generalized business model of contracting out to third-party vendors. When you go the supermarket, it's very likely that the cash registers came from IBM (at least in the US). When you go to a restaurant, your payments are likely to be handled by Authorize.net (not to mention the credit networks), when you go to a hotel, the wifi is handled by some third-party company.<p>In each of these cases, there is a valuable service being provided to a company, and any service that has a large enough market penetration has access to a large amount of information about you. Third-party vendor relationships are not going to go away.<p>So what does this mean for the net? What effect will this have on the world? In the best case for anyone who thinks this is a good idea, it won't catch on, doesn't go mainstream, and a small amount of invisibility is granted to the few users who adopt this. But what happens if this were to become the default way the web works? Think about the wrench/security XKCD ( <a href="https://xkcd.com/538/" rel="nofollow">https://xkcd.com/538/</a> ). If browsers started doing this, than the companies who depend on these services will just find other solutions. Already, many sites use url redirectors in order to accomplish this (think: Google Search), if we try to prevent redirects, then you break the web.<p>Okay, nuclear option, we decide to break the web, and require all users to manually confirm before doing redirects. There will be sites concerned enough about the drops in traffic that would be caused by forcing redirects that they'll stop using them. Does that mean this practice will stop?<p>Of course not! The next step is out-of-band network requests. You request data from a site, they ping the third-party, no muss, no fuss, and you're never the wiser. Well, what happens in that case? Suddenly, you go from a situation where companies are using third-party services via the browser, where the end user's interface to the third-party is entirely secured from the company receiving the service, to a world where all of this information is proxied the third party, meaning that instead of less companies being able to track you, suddenly more companies are able to track you.<p>I for one, hope that this does not catch on. Not because I care one way or the other about the tracking (I think there are better solutions<i>) but because I think this is an effort that will hurt both the web, and eventually users.<p></i> My thoughts on this: I find this akin to state surveillance. It's inevitable. The 'solution' to state surveillance is not hoodies and masks, but instead sousveillance (which means granting distributed surveillance powers to the masses, like way a large number of police issues came to light). Put the power in the hands of the people. How does this apply to web traffic? Same thing, instead of having traffic available to a few companies, make it scattershot, and make sure everyone has access. Surfing the web? Make browsers overfetch. DNS lookup for the letter p? grab everything from park to porn to production. That way, there is no loss of power