Earlier quoted context omitted.
Did some digging, here is the paper he references on echo chambers/bubbles: https://journals.sagepub.com/doi/full/10.1177/21582440198327...
My initial concern, based on reading the abstract, is that since this study only looks at Twitter and not Facebook, it isn’t enough evidence to make any conclusions about echo chambers in a general way as so many people get their news from Facebook.
Reddit popular is a total echo chamber. Almost every sub, outside of focused hobby subs, turn into echo chambers eventually.
The dynamics of reddit, with downvotes totally squelching dissenting opinions and upvotes promoting popular ones, is naturally going to turn into a total echo chamber.