I would argue that things like the massive photo collections people have made provide a large level of lock-in. I also suspect that Timeline is the first play in a "all your history, here" move that would make moving to another service very difficult.
But I don't use YouTube any more. The level of advertising got to beyond tolerable.
And that is the problem these sites have. I might spend four hours on the weekend with music videos on, but I just could not increase that, there is not enough hours in the day. To grow, the site has to increase the proportion of ads, and then it starts to lose users based on their level of tolerance for crappy, repetitive advertising.
So anyone buying Facebook and suchlike for growth potential should also consider that there really is a practical upper limit to the revenue it can pull in before it becomes annoying and starts losing users. IMHO, YouTube is beyond that limit now, Google is pretty close, Facebook has a little room, but not that much.
This can be excessive ads, but this could be as well things like excessive features which make the service more bloated and unsuitable for the original audience.
Zuck could do shit, because that downloadable app would use client local IP, so they would have to block _everyone_ in order to stop the app. Further, it would use built-in IE frame, so all request would look like coming from legitimate Internet Explorer user. You could also implement "wait(rand [sec 1-5])" between downloading requests, so it looks more messy, more like a human that is clicking, rather then robot pulling requests in miliseconds.
Zuck once had hacked into school system. Who knows -- would be crazy to see an App that so easily ported out all your data out of Facebook, and see Zuck trying to stop it :)
If you think you could download people's entire photo collections in a "couple of minutes" then you underestimate the insane number of photos uploaded to Facebook every day.
What I said is that the App would download all _your_ photos that you owe on your account. Given that it takes couple seconds max until I see a full photo from the moment I clicked it, it would take tens of minutes to download hundreds of photos: 5 sec per photo x 200 photos == 16 minutes of robot's work while you enjoy your drink.
edit: so this "I would argue that things like the massive photo collections people have made provide a large level of lock-in" becomes an irrelevant problem.
is it really that hard to imagine??