Not Recommended is an experimental anti-recommendation engine that provides suggestions for things you probably won't want to read, watch, or do. It is an attempt to mitigate the effects of filter bubble by appropriating an existing and familiar web technology in an unexpected way.
The site uses Amazon Mechanical Turk to generate anti-recommendations based on data from your Facebook profile. I asked anonymous workers to provide “anti-recommendations” for about 1000 common books, movies, TV shows, and interests. When you log in, the site checks your Facebook likes to see if it contains any of these items, and if so, displays the anti-recommendations. If your Facebook data is too sparse, you are prompted to select from a list of six items for your favorite, and then you are given the anti-recommendation for that particular item.