Medium Feeds, myopia and a possible cure
The more I read my Medium feed, the less aware I become. It is like quicksand — every attempt to extract myself only entrenches me further in one way or another.
Are artificially intelligent ‘recommended’ feeds promoting a new myopia? I reflect on recent experiences and make a case for ‘reverse-curation’.
Paraphrasing from Merriam-Webster dictionary, myopia can be defined as:
1) a condition of the eye that makes it difficult to see objects that are far away
2) a narrow view of something
Myopia is a ‘refractive’ error of the eye which means that “light is not bending properly when it passes through the lens of your eye”.
This refractive error can be literal or figurative. With automated feeds, it is the latter. I use Medium as an example that readers can easily relate to. There have been a few posts like this one Medium feeds.
While the exact algorithm is not known to me, a recent post by Medium indicates that content preferences are driven by publications, tags and users we follow.
Personalized feeds feel like a ‘walled garden’ where the boundaries seem finite and not as malleable as I’d like them to be. For me a big issue is that I want to read widely than the recommended interests/tags. This is hard because the feed mostly ties us down to topics anticipated from recent activity. The ‘For You’ feed seems to be based purely on personalized content. In fact, I also group ‘Top Stories’ into a similar category because if we are following popular tags, then there will be overlap. Also when a big event is happening, the top stories will tend to be on the same topic — another paradoxical case of where more is less.
I’ve tried many ways to broaden my perspective. This includes un-following some people, changing feeds to “show less of” stories, and even un-liking some posts that I like so that I don’t see more of an author whose post I liked once but whose subsequent posts are either too repetitive or irrelevant.
These ad-hoc corrections create two problems. Firstly I have made so many changes that it is now impossible for me to judge how ‘close’ or ‘far’ away my feed is from my ideal feed or indeed even the previous ‘version’ of my personalized feed. There is no ‘baseline’ so ad hoc changes are only changing things further. This blind pruning is rather dis-orientating.
Secondly, I’m now terrified to even like a feed even when I like it because I often preview the implications — what if I like this and then I (only) keep on getting more things like this — which I don’t want? Wonder what net impact liking this or comment on this post would have on my feed that I’ve worked hard to tweak?
Since I’m the end user, my only ‘control’ mechanism is ‘adjusting’ the feeds. The problem is that I’m not sure how much better or worse my ‘adjustments’ are making my feed. Is the myopia we’ve talked about being reduced or merely being replaced by another myopia?
If we go back the eye metaphor, these tactical measures can be described as a coping mechanism — like squinting to more clearly see distant objects.
What we need is cure.
Myopia can be corrected by lenses — concave lenses. Notably, these are diverging lenses and that is precisely the cure required — divergence.
What I would like is a control panel where we can set the level of customized content we want to receive. The closest analogy I can give is like preparing tea. The core ingredients are tea (flavour of choice) and hot water. Milk and sugar are ‘added to taste’.
It would be great if we could have a ‘tuning bar’ that allows us to directly control how much or how little customized content we see. Alternatively, different customized feeds (publications, users, tags) can be presented as ‘channels’ that appear in addition to the normal ones (e.g. Reading Roulette, Editor’s Picks etc.) in Medium.
What I would especially like is a “Not For You” feed that focuses on content from writers who have no interests/tags common to my readings. This ‘reverse-curation’ will allow us to read widely and encourage serendipity. Actually, we can even build in a tag-analysis mechanism that tells people how closely clustered their tags are within the wider spectrum?
Paradoxically, what I would also like is a ‘Mute’ button of some description. This is to mute people who write ‘click-bait’ or simply bombard topics with populist topics or one-sided ideological perspectives. In effect, a sort of noise filter, that is directly controllable by users, would also allow us to better see the forest for the trees.
Constant near vision work gradually diminishes our ability to see further. Personalized content has the same risk profile.
When I look at the “What is Medium” post in the Help section, the key words jump out as Write Freely and Read Actively. Yes, I can read ‘actively’ but I want the ability to read widely. If I can Write Freely then surely a counterpart should be to ‘Read Widely’.
Designing for a wider reading perspective and serendipity would be welcome and perhaps crucial to combat the new age myopia of personalized feeds.