I think Eleanor Saitta’s maxim, “all technical problems of sufficient scope or impact are actually political problems first”, applies to this case.

infosec.exchange/@dymaxion/109…

(continues)

Zoom out sufficiently, and you will find that the problems of this tiny social medium stem from intersecting dimensions of colonialism, imperialism, capitalism, and all other forms of exploitation.

@weaver touched on some of these aspects in their posts.

zirk.us/@weaver/11403078544891…

(continues)

So... you cannot build an egalitarian social medium in a world shaped by exploitation and exclusion. The social medium will reflect those inequalities and biases, they won’t magically disappear in the virtual world.

(continues)

Nevertheless, if you, dear reader, want to make the #Fediverse egalitarian, go for it. But keep the larger context in mind, that will empower you to do a good job.

(continues)

I hear you think: do you have anything concrete to suggest, you pompous ass?

Yes, if you insist... 😆

(continues)

I think the #Fediverse should first solve the problem of search and (universal) discoverability, even at the risk of creating potential point(s) of centralization.

Next, the Fediverse should introduce the dreaded “algorithm”, a built-in mechanism to provide a custom feed to each user based on their ip address, preferences, recent posting history, and whatever else (preferences and recent posting history of connections?) usually goes into those magic formulas.

(continues)

The “algorithm” must be transparent, and must be community-owned and community-controlled [which calls for some forms of (messy) governance structures]. The “algorithm” will depend on user-specific data, so there will have to be rules/regulations/terms and conditions/(agreements?) regarding data collection, retention, and deletion (say, nothing older than three months should be retained, or used by the “algorithm”).

(continues)

Whether users get to customize the “algorithm” for their feed, whether it is opt-in(out), or baked into the software—all these are details to be worked out.

(continues)

Will an “algorithm” increase the resource load on servers, making it harder for small servers to continue their operations? Maybe... I don’t really know.

I think the “algorithm” should be designed with this concern in mind, it should be made as light on resources as possible.

(continues)

I have said “algorithm” in the above posts, and will keep using the singular in subsequent posts, but there may also be a plurality of “algorithms” that do the same job slightly differently, and users or administrators may get to choose the one they prefer.

(continues)

(I know all the talk about an “algorithm” processing data sounds alarming, especially to a Fedi user, but we should keep in mind that all the #Fediverse servers have varying levels of data stored right at this moment, with varying retention policies and varying levels of compliance. And almost all of the Fedi is public, including the connections between users, so search engine scrapers and AI crawlers are indexing practically all the data all the time.)

(continues)

In my opinion, only the implementation of these two steps: near-universal discoverability of users/topics, followed by an algorithmic feed customised per user, will enable low visibility users/topics to meaningfully participate in the #Fediverse.

(continues)

Hardly any user has the privilege to spend hours wrangling with lists and hashtags and other paraphernalia to find a couple of engagement-worthy posts.

And as my thread on “global south invisiblization” explained, all that wrangling cannot overcome the erasure of low visibility users/topics; so what’s the point?

I believe that the two steps I outlined will significantly increase the number and the diversity of users/topics on the #Fediverse.

(continues)

At the same time, I know that “no algorithm” has become a flex for many Fedi users.

(continues)