From the number of squares, I thought it might end up doing something even more 'clever' i.e. generate a square for each of the most recent n URLs from feeds of

From the number of squares, I thought it might end up doing something even more 'clever' i.e. generate a square for each of the most recent n URLs from feeds of m news sites, then analyse, for example, words in headlines of those articles to determine what I'm interested in. Lots of potential for data analysis once you have someone's browser history.