~1y

Sometimes Algorithms Help Us

"We Value Humans Over Algorithms" writes @budnitz in his latest welcome to newcomers to Ello ... and I agree that algorithms that manipulate what you see are bad...

BUT...

... that doesn't mean that ALL algorithms are bad.

Back in the early 2000's when "blogging" was so new it was wonderfully easy to follow so many other bloggers. There was a small pool of those of us who blogged about voice-over-IP (VoIP) and we could easily all follow each other, read each other's posts, comment on the posts, subscribe to each other's feeds, etc.

It was a smaller world.

But as the "blogging revolution" unfolded and more and more people started to write online it became harder and harder to follow all the blogs on a topic. Soon you needed an RSS news reader and some mechanism for sifting or sorting through the data. Soon you HAD to use tools and algorithms to help you find the content most interesting to you.

The deluge of content became too hard for one person to handle.

When I joined Twitter in October 2006 (danyork) it was wonderfully easy to follow so many other people on Twitter. There was a small pool of us who were trying it out (I was Twitter user 10,000-and-something) and within our circles we could easily all follow each other, read each other's tweets, reply to them, etc.

It was a smaller world.

But as the masses flocked to Twitter and more and more people started to tweet, it became harder and harder to follow everyone of interest. Soon you needed to use some kind of mechanism for sifting or sorting through the data. Perhaps you followed just a hashtag ... or followed a search on a specific term. For me it was Tweetdeck with a range of columns targeting different terms, tags and people. You HAD to use tools and algorithms to find the content most interesting to you. Today I pretty much NEVER look at my raw Twitter feed... over the years I've wound up following far too many people and accounts.

The deluge of content became too hard for one person to handle.

When I joined Facebook in 2007, the experience was similar... the big difference is that Facebook realized along the way that this evolution would occur and took steps with the NewsFeed to address it. They realized that people needed tools to handle the deluge of content. They started using algorithms to "help" people find the content that mattered most.

And if you think back to the early days of the NewsFeed, Facebook provided all sorts of controls to users to tweak how the NewsFeed algorithm worked. I seem to remember a whole series of sliders that we could adjust to show how much we wanted to see for various topics and themes. Those of us who like to play with options like that had a lot of choice.

Somewhere along the way, though, I seem to recall that Facebook decided that all of these options were too confusing to many "regular" users... and so they removed them. A more cynical interpretation can be, of course, that they wanted to control and manipulate the feed themselves - and to also figure out how to work in more sponsored content.

Regardless, we're now where we are today with a NewsFeed algorithm that is opaque ... it's a "black box" that Facebook can (and does) change at their whim. They can manipulate what users see ... and therefore have a scary degree of control as a gatekeeper of content.

And now we hear that Twitter will be introducing more algorithmic control, too. They've already been doing a degree of this with the "While You Were Away" ... which, quite honestly, I've found hugely helpful on the few times that I go to my main Twitter home page.

I think the issue with the huge anti-algorithm reaction on Twitter is because the expectation is there that it will be an opaque "black box" algorithm like Facebook's algorithm is. And I'm sure it probably will when they eventually roll it out.

At some point in its growth, ELLO WILL NEED ALGORITHMS in order for Elloers to survive the deluge of content.

It will.

Or else people will find that they can't deal with the volume of content here... and may reduce their usage and/or move to some other service with a smaller pool of users to start the process all over again.

The key issue is transparency and control of the algorithm.

Back when I was using tools to manage the flood of blog posts, I was in control of the algorithms that I used in the newsreaders and tools. When I currently use Tweetdeck to manage my view of Twitter, I am in control of the algorithms. I set up the search terms and the columns.

The issue with Facebook's algorithm and that which we assume will come from Twitter is that we as users have no clue about how the algorithms work and also have no way to influence those algorithms.

This will be the challenge for Ello - how to introduce algorithms that can help Elloers navigate the volume of content, while allowing the users to be in control.

Oh... while still having the simple, minimalist feel of Ello...

I don't envy the task that @budnitz, @cacheflowe, @lucian, @todd and the others will have when Ello reaches that stage of its growth. (Which, on one level, will be a great problem to have!)

My point now, though, is this... try not to demonize algorithms because at some future time Ello may very much need some kind of algorithm to help.