I've never really been very big on the 'filter bubble' theory. Whilst I can appreciate the thought that increasing personalisation of digital services based on data can shape our input in some areas I think it's too much of a generalisation to say, as Eli Pariser does, that we live in our own personal universe of information online and that "the internet is showing us what we want to see, but not necessarily what we need to see". I think the variety of services we use, the variable methods of discovery, different combinations of curation, and the choices we deliberately make to broaden input act against this.
If I think about my own consumption habits, the breadth of sources that I consume now (both passively and actively) has exploded in comparison to ten years ago. Quite apart from anything else, the effortless globalisation of content means that I'm just as likely to discover and read something from The Atlantic, or the Sydney Morning Herald or Campaign Asia or Aeon as I am from The Daily Telegraph, and the same is true of the blogs I read, the videos I watch, and the content I happen across by falling down those link rabbit holes that we all do. A new study by Newsworks and Twitter has found that 60% of respondents had engaged with newsbrands on Twitter that they wouldn't typically read in print. I think it is too simplistic to assume that increasing sophistication of algorithmic curation in some services automatically means a deteriation in overall serendipity and diversity.