On ideal feeds, and a defense of The Algorithm

19 Jan, 2025

I've had much rumination over the past few days on the value of personal, chronological feeds versus the modern algorithmic feed driven by engagement and capitalist incentives. And in general, what makes an ideal feed?

Before I continue, I'd first wish to clarify that for the purpose of our discussion in this post, "feeds" are streams of "content" created by people presented to consumers. "Content" can be textual or media. In more broader terms, I'm talking about the usual feeds we get from social media apps, RSS feeds, news sources, magazine issues, and so on.

Like most things in life, I can already hypothesize that there is no one-size-fits-all and that the best we can do is seek a balance and a "good enough". However, for the sake of organizing my thoughts I've decided to list down all the kinds of feeds I could think of and try classifying them to see if there is any pattern or conclusion I can draw.

Different people have different answers. Here's how I would go about it.

Classic "following"

These days, I'm only on social media for the purpose of knowing what my friends are up to. In plainer terms, I wish to see whatever content published by the specific list people. Of course, the feeds of modern social media don't end up being this way. But it's certainly what some of us stay on social media for, and what I want out of social media today.

Using RSS/Atom feed readers to follow personal blogs can be classified under this kind of feed: the type of content I get is determined entirely from the people I follow. Everything I get is made by those people.

Personal curations and bookmark lists

Some people publish a chronological list of cool things they have found made by other people. Examples include feeds of "starred articles" from their own feeds.

The content I get can be from anyone. It consists of anything that the specific list of people wants their readers to see.

In this case, the judgement for whether a piece of content comes my way rests entirely on the specific list of people whose bookmark lists I'm following.

Zines curated by a single person can be put under this category.

Grouped/Community compilations

Everything from members of a group with no further algorithmic filtering. Following a feed of this type means I want to receive everything and anything made by anyone part of this group. Content is determined by whether its author is a member of a specific community.

Example: the "New" filter on Reddit communities.

Publications

This is essentially the classic following list, only the type of content I receive is curated to be in the best interest of the group of people, organization, or company behind the publication.

All content is made by a specific list of people part of the publishing organization. I see whatever the organization wants me to see.

Zines curated by multiple people can be classified here.

Algorithms and "The Algorithm"

Here, the content can be made by anyone, from anywhere, so long as it fits certain criteria determined by an algorithm. The algorithm may serve the best interests of particular individuals, or particular things.

Examples include:

• Bluesky Feeds that might comprise all posts of a given topic, posts that contain a certain set of emoji, etc.
• "Trending"/"Rising"/"Controversial" threads from forums or link-aggregators. In this case, all content is made by members of the specific community. Whether something is included in the feed, and in what order, might be determined based on the level of engagement a post receives by the community members.
• Personalized recommendation in modern social media. The algorithm is tuned to drive engagement and revenue. I might get the usual interesting content, or rage bait, opposing views, viral media, or anything in between.

There is no ideal feed

After breaking it all down, it's become clear to me that there is no obvious winner, nor loser.

For those who consume feeds for the purpose of entertainment and occasionally serendipitous discovery, I can't seem to argue purely for personal feeds against the algorithmic feed -- as is the trend of the population that defend an open web and against modern social media.

Letting a few individuals of your selection determine the kind of content you see doesn't seem to be such a good idea.

The best way to go, then, is to have a mix of everything. But why? And is it really productive to make these clear distinctions between the kinds of feeds like I did above regarding the author of the content and how judgement is passed?

Everything is an algorithm

I eventually came up with this theory in the attempt to find the most satisfy explanation.

All the kinds of feeds I've listed are filters and sorters for content.

Consider a fire-hose of all content produced in chronological order of a given scope -- say, on the entire internet anywhere, anytime. Feeds, then, determine the selection of content I get to see and the order I see them in. By subscribing to RSS feeds of personal blogs, I am subscribing to particular individuals' interests and the type of content they usually publish, and the order they publish it in. I can subscribe to zines to receive content as curated by the publishers, in the form deemed appropriate by the same, for the zine.

Trending content may selected and sorted by metrics, calculations, and randomly selected multipliers rather than real humans, but our thinking and judgement is only a complex algorithm that can also be influenced by chance.

At the end of the day, they all filter and sort content in some way. With account blocking and keyword-mutes, I am further tailoring the filter to my needs.

When I apply filters by article tags to RSS feeds, for instance, the algorithm becomes "anything tagged with X and Y, without the Z tag, published by this person with whatever extra heuristics this person takes into account when publishing their content".

What we usually consider to be algorithmic is only more complex versions of this with dependency on more factors. This, of course, includes chance.

It's worth noting that for the purpose of our discussion in this section, an "algorithm" need not be specifically limited to specific calculations and formulae conjured by real humans. It is simply a means of selection.

As an author, I might choose to publish this and discard that draft simply because the stars aligned tonight, or the wind is blowing north-east. Likewise, it's often impossible to predict the outcome of a social media posting having passed the judgement of The Algorithm¹ and the waves of engagement and impressions from the millions of users on a platform.

By following multiple feeds of different kinds, I am following multiple algorithms.

Closing thoughts

These conclusions are drawn when I only consider consumers of given feeds. The story might be very different if I take into account the impact of different algorithms to the content creators, and the feedback loop it forms.

Much also remains to be said about the way we approach the content to be consumed even after the filters and sorting of algorithms, and how it influences our perception of the value of our feeds. For example, what difference does a "read"/"unread" indicator make? Compare newsletter issues that arrive at our inboxes as unread emails and the need to archive or delete that email, with an infinite-scroll homepage of a social media app where a simple scroll-up gesture is enough to forever ignore and forget about a recommendation.

What remains important is recognizing as consumers that we have a choice in the kinds of feeds we subscribe to². Rather than succumbing to consuming whatever The Algorithm sends our way, or -- at the other end of the spectrum -- refusing to be swayed by public impressions and the best interests of entities we cannot control, we should recognize that one isn't all so different from another and the best feed is combinations of one we select from multiple sources each with its own worth and intrinsic value.

Reply via email