I don't think you browse like I do

Those bastards... and they force you to sign in for their email and g+ and picasa and.. oh, wait.

Maybe I'll sound like a wierdo, but you can always opt out of using google at all and block them with ghostery or hosts. Problem solved. Let's be honest with ourselves and admit that data is the currency we pay for 'free' online services, and if we don't like the service or the 'price', we can just stop using it.

It's funny OP uses gmail and g+ but is worried about profiling for ad personalization. Btw there is google analytics on his blog.

This is FUD, plain and simple. While I don't advocate sticking one's head in the sand and refusing to predict or even speculate, it's easy to do too much predicting and speculating (and very prone to bias), it's easy to spread paranoia, and it's no basis for living a life and especially it's no way to decide moral or criminal issues. I could go on a murderous rampage tomorrow (perhaps not physically possible as determined by the laws of physics making up my brain but "possible" in the looser sense of it being a member of the state-space of world-configurations). Why don't people I know think this is likely? Largely due to trust that I've built up. Someone once asked me what they could do to gain my trust back when they broke it, I said that it's largely a matter of not doing certain things for a long enough period of time.

And so the problem is heavily dependent on what Google, Facebook, et al. end up actually doing with the data. What they could or might do has little relevance, what they're likely to do has significant relevance but still not as much as what they actually do. Thus far, they've been awesome about either not doing evil things with it or at least not letting me know about their evildoings. Their trust is not yet broken with me, unlike with some other companies I can think of (and correspondingly don't deal with at all).

Therefore I continue to use their products and let them collect and infer as they please. I don't exactly make this easy for them either since I too have "unusual" browsing habits in the form of NoScript, AdBlock, multiple browsers, shared, dynamic IPs and occasionally Tor. I didn't even join Facebook until a couple months ago on a "Let's see who friends me" curiosity (so I don't friend anyone myself). I'm sure there are plenty of Evercookies on my system I'm unaware of though and my full Firefox UserAgent is pretty unique. So they likely have a decent profile of me that they just don't reveal in a straightforward manner. They could also form a good picture from my comments here and elsewhere. (I don't think they have much ad preference data on me except perhaps that I hate ads.)

Though that brings me again to the issue that it can be worth speculating over what is likely to happen that will cause the trust to be broken and if I should be avoiding certain classes of actions that typically lead to harsh repercussions later. For example, suppose you have data showing it's quite likely that lending large sums of money to friends leads to the destruction of the friendship sometime later. Maybe your friendship will be different, but if that outcome is quite likely, you'd probably be better off not taking the risk and in this case acting on facts apart from what the other person has actually done is useful.

Even if these companies break their trust with me in the future, the most likely ways it will be broken do not seem to have incredibly unfortunate outcomes for me. (I can of course think of other quite nasty outcomes as well as anyone.) Similar to how, for example, a friend stealing something petty from you doesn't often cause great financial loss or inconvenience but the trust is shattered nevertheless. Or there may even be actions taken that I never signed up for or wanted and probably would have said "This will violate my trust if they do this" had 'this' been told to me before they did it, yet those actions actually end up being very useful to me and so it doesn't break my trust. I think this is actually how a lot of technological progress has been for a lot of people. They're taken on the ride from the status quo whether they like it or not, but in the aftermath they often like the new place better.

The paranoia that some company might understand what i want and try to provide it for me seems absurd. why is this such a bad thing?

>I just prefer web sites that aren’t updating state based on me just reading things.

That seems odd. What do you care? Do you get frustrated if they incremented a counter to see how many people are currently on their site?

>I like to see opposing sides of the argument, so don’t like the idea of being bubbled by my own preferences.

That's shortsighted of you. The algorithm can detect that you like this and will show you a healthy does of opposing arguments.

>I feel a little freaked out when one cycling site I go on has adverts for the tyres I looked at 2 weeks ago on an entirely unconnected web site.

So you don't like to see relevant ads.

>Personalisation can get it wrong. In the same way that Amazon recommendations become tainted every December when you do the Christmas shop.

Sure, it can get it wrong, that's why it isn't as prevalent as you make it seem to be. Also, Amazon does their recommendations based on the latest things you shopped for. So when you start browsing other stuff, they'll show you related content.

>And personalisation can be extremely upsetting, such as a family member who still receives new born baby information months after a miscarriage.

Any one can get that wrong (and this isn't an extreme case). Their friend could congratulate them on the baby.

>It just comes back to how I use the web.

Like old media that isn't dynamic.

>I enjoy it more when it’s made up of many disconnected things giving a consistent experience to the user... me.

That doesn't make sense. It seems you don't like the Internet.

You make some good points on how they could improve these things, but not on them not being used.

I din't realize inaccurate personalization could be so upsetting. I typically take everything I read or see with a grain of salt. The Internet, afterall, is made by bozos like you and me ;)

I sometimes dream of a browser that would just fetch the page content and no extra other resources, showing it like Instapaper. The reality is more complex: images, css, CDN's, dynamic content, ...

RequestPolicy realizes the most important part of that dream: not telling Google, Facebook and a host of obscure tracking companies about every website I visit.