Every time we bookmark a website we not only save it for later but add a piece of information to the page itself. Each bookmark can be treated as a 'Like' and every folder can be represented as a Tag. All that information can be used to extend the meaning of a bookmark as it was to be - a collection of urls in menu bar of the browser.
I don't stop there, and I don't store only links to websites. At some point I started using the bookmarking services as a tool to add annotations to websites. From simply tagging videos I watch on YouTube, rating movies I see on IMDB to saving quotes from articles I want to reference in the future. This way I don't have to create yet another account or Like something on Facebook, and the data stays free and open.
To be honest I don't really know how to know this type of usage so I will use an example. During my thesis at CIID I decided to collect reference materials and documentation of the process. The problem was that they were all scattered in several online services: images at vi.sualize.us, links at Delicious and photos at PicasaWeb. Because I wanted my thesis advisor to be able to track my process I built a small script called stream that pulls items tagged with ciidthesis label and displays all of them on one page. This way ciidthesis tag become something more than just a binder holding items from one service together but also a cross link between different services.
In the examples on the top of this post and below I used data from url lookup page that lists all the bookmarks of http://marcinignac.com to see which tags people use to describe my website. You can click any of the images to see SVG version of the image along with the code.