Feedback loops

Communities are often accused of being feedback loops. We naturally gravitate towards things, ideas, people that support or confirm our opinions.

Filter bubbles are blamed for divisions in opinion and society, just look at what happened during the US election and the run up to Brexit.

Being a Google property, Google Plus has long been considered a hotbed of anti-Apple (or, at least, anti-iOS) rhetoric where Android can do no wrong.

It's only natural.

Highly engaged and devoted Google Plus users are far more likely to have Android phones and people get very passionate about their platform of choice.

For some reason I clicked on the link to see the all time most popular posts on my blog. The top three posts, and by some margin, all had something in common:

  • they were all about Google Plus, and
  • they were all shared on Google Plus

Communities just love reading and talking about themselves and the platforms they run on.

All social networks go through a "meta" stage when the biggest topic of conversation is the network itself. We expect this while everyone gets used to a new way of doing things and deciding, as a collective, if that way is adequate or can be improved.

But my example above really brought home just how self-reinforcing these things can be. For these three posts to be the most "popular" purely because of a loop effect is... Worrying? Disappointing?

I'm not exactly sure of the emotion or the words I need to describe it, but I know that it's not the way things should be.

Feedback loops

Can we ever make the Medium we want?

Forcing change.

Maybe the algorithms have become more responsive or maybe I have changed the way I use Medium and not previously noticed how responsive they can be.

Follow new people, recommend different posts in different tags and a host of new content based on those actions instantly appears in my feed.

It has become a lot more obvious how certain aspects of the algorithms work and how we can use them to "fix" our feeds. But there are still gaps, little black holes where content gets sucked in and we don't exactly know to what degree.

"Show fewer stories like this" sounds like a good way to remove unwanted clutter but how does it account for nuance? Where is the event horizon beyond which nothing returns, nothing is visible?

But there are also crazy wormholes, story spewing singularities that respond to our actions in ways we would rather they didn't.

A follow or recommend will immediately give us more from that person but there seems to be no flow control, no finesse. The algorithms get over zealous, regurgitating unrelated stories from years ago just because they are by the same author.

It's all or nothing.

This is the problem with follow and block, with algorithms that work on rules - I like posts about Twitter, I don't like listicles.

Include and exclude.

What happens when those rules intersect and clash? There will always be exceptions so what takes priority?

We are creatures of habit, sure, but "something about us compels us to learn, explore" and an algorithm cannot match that compulsion, that curiosity, that nuance of thought and taste.

How many of us leave the safety of our feeds and jump with both feet into Reading Roulette or scan the trending topics?

Despite our best efforts and digital snobbery, we run the risk of getting caught up in the same filters bubbles we so detest on Facebook. We constantly need to keep things fresh but how much effort do we, should we, expend in tweaking the algorithms?

Is it even possible or are we wasting time pursuing an unobtainable facsimile?

Are we forever chasing the white rabbit?

Can we ever make the Medium we want?

Filter bubbles in an echo chamber

Much has been written about the danger of filter bubbles and echo chambers over the years, especially with regards to the social web.

New research into political news consumption indicates that most people actually use centrist, mainstream sources rather than those at either extreme.

So what's the worry?

Although most would appear not to live in echo chambers, media consumption is widely skewed and those who do are disproportionately influential.

Birds of a feather

The social web is a different beast and it is widely recognised that we follow accounts and sources which are sympathetic to our own views rather than subject ourselves to confrontation - although it often comes looking for us.

I have deliberately made a point of following some accounts that I probably normally wouldn't based on my every day interests but recent events have made me consider how much further I need to go.

The "chosen tribe" comment was unfounded but, on reflection, I can see why it might have been made based on the subject matter at hand and who was being referenced during the conversation.

But it was cleaning up my RSS feeds that really got me thinking.

It wasn't so much reducing the feeds I follow from over 200 to just 51 but the realisation that over 75% of the feeds I had subscribed to were no longer active - this meant that the hundreds of items I review each week were coming from such a small number of sources and I didn't even realise.

On Twitter I follow less than 300 accounts. This has risen recently as I have tried to broaden what I am exposed to, but I can't help think that this is still too small a number and some of those may be dead accounts.

(Update: I went through those I follow and removed around 15 so it wasn't as bad as I expected.)

On Medium I follow a wider variety of people, largely due to the nature of the platform and my reasons for using it, but I still only follow 227 people - a list which I have never sanitised.

The problem

Discovery is a massive issue on social networks, as is pulling signal from the noise. Because we don't want a feed full of irrelevance it is only natural that we should follow those who provide information aligned to our interests.

It's not necessarily that we go out of our way to avoid others but that we are making snap, subconscious trade-offs between the value we derive against the potential for noise that may be created.

It is an inherent issue with following other people: because we are multifaceted, no matter how much our interests align, there will always be an element of irrelevance.

Without meaning to, and perhaps more importantly without even realising it, we inherit a narrow world view which gets further entrenched as we seek to eradicate the noise.


Noise disturbs and distracts us, makes us uncomfortable, and ruins our experience. I'm not just talking about social networks but the parallel is perfectly applied.

In life we will cross the street, move to another carriage on the train or turn our headphones up a couple of notches, retreating to our own little bubble.

Online we unfollow and block, maybe even move to another social network. Anything to escape the noise.

But we need the noise, we need to be pulled from our reverie and made aware of what's around us, no matter how uncomfortable it might make us feel.


And so we return to the problem of discovery - not just the actual finding but the self-discipline we need to pull ourselves away from the safety of our filter bubbles.

We may flick to trending topics or browse Twitter Moments but these are merely sources of brief, instant gratification - a quick fix before returning to our regular programming.

The question becomes how do we as individuals locate contrary, challenging content without the conflict that so often arises with it?

Social networks are so focused on trying to show us what they feel we might like but there needs to be a way to surprise us, to throw us a curve ball and stop us in our tracks.

There needs to be a way to force us to think and reconsider, but it has to be subtle - that sounds like an oxymoron. It cannot be so forced upon us as to make us recoil and reject it out of hand.


It is contrary to almost everything the networks are built on and how we are wired.

Filter bubbles in an echo chamber