First off, for the hopeless pedantics among you, let’s get this out of the way: yes, I am aware of the definition of “algorithm,” and I realize that literally any arrangement of feed items into a feed is algorithmic, even a chronological feed. But “algorithmic feeds” has a clear colloquial meaning here and I will hereafter use the term “algorithmic” to refer to feeds or timelines where the content isn’t based strictly on straightforward things like who you’re following, but rather a bunch of complex choices, usually optimized around maximizing engagement.
I keep hearing people running into literally any friction whatsoever bootstrapping their community on Mastodon and their immediate suggestion is that Mastodon needs the algorithmic timelines that all the other major social networks use, and that Mastodon’s failure to do so is relegating it to failure.
There’s a lot about this that’s a bad idea but it’s all predicated on this fundamental misunderstanding that the algorithmic timeline exists to benefit you, the user.
And sure, that’s how algorithmic timelines get sold to you, but the real reason tech companies have invested so much in making timelines work like this is because they want growth and don’t care about anything else and will go to ridiculous lengths to fill feeds with content that are statistically more likely to make the app “sticky,” keeping you on there for longer. It’s not because big tech companies believe in earnest that you spending more time using their stuff is better for you; it’s just a nihilistic mentality that growth is good. It’s an empty belief system and it’s naive to think it’s better for you. I like that Mastodon isn’t burning through millions of dollars VC investors’ money a year desperately trying to increase growth without bound.
It is an important philosophical difference between Mastodon and mainstream social networks that you have agency over the people you see content from in your feed. When you cede that control, you’re giving someone else decision-making power over your media diet. You’re giving someone else control over the things you learn, and how you grow.
But the thing that really just pisses me off about the “maybe we need an algorithmic timeline” takes is this false dichotomy that the only two options are how Mastodon activity feeds currently work, and how Twitter’s activity feeds now work.
We spent years lamenting the consequences of the algorithmic timelines. We talked about how they created echo chambers and sowed more political division, and when faced with obstacles when using Mastodon with chronological timelines our immediate knee-jerk reaction is “let’s go back”?
Fuck that noise.
There exist other solutions that can help address the issues people want to solve with algorithmic timelines.
If seeing every post in your chronological timeline is overwhelming, we could experiment with other ways of presenting those posts. Maybe if you’ve got some serial posters who clog your timeline, we could collapse the posts automatically. We could experiment with letting you look at a digest view of your timeline, perhaps letting you filter to just posts that are popular or getting the most replies. Hell, we could even experiment with LLMs that try to summarize your feed content or sort posts by topic. There are lots of filtering options not yet explored that can make it easier to sift through a timeline full of content while still keeping the philosophical spirit of control over who is in your timeline.
Mastodon has a “reply guy” problem, and as it’s getting some notable users they’ve been raising alarms about it. I would like to see instance admins moderate reply guys better (seriously, if you’re a Mastodon admin, it’s okay to discipline users for just being assholes in replies to people; it’s not a human right to be able to antagonize people on Mastodon). But we could also offer more robust options to limit replies, or to display lower-quality replies separately (which is flirting with “algorithmic timeline” in similarity, yes, but we could do it in so many ways). There are so many different options for heuristics that could be used to help with that issue.
And yes, a fresh and brand new Mastodon account is daunting. And that is a tough problem so solve automatically without resorting to things that look close to algorithmic feeds, and an account’s beginning is an important time in its lifecycle because the people you initially follow serve as the seed for the community you eventually build. But this is solvable. Human curation is an option. I’d suggest using contact lists to help build social graphs but I know so many social networks abused the crap out of this and it feels like a bad idea. But we can build better search tools, and clients can help you find your people by asking a handful of questions early on that might help you connect to the right people.
Plenty of people are more stubborn about Mastodon’s need to evolve and add new features; its growth reveals that its needs are different than they were in Mastodon’s early years. Mastodon has handled its growth with grace, and it continues to look a lot like Twitter did in the early 2010s. Now is a chance for Mastodon to start to take a path that is uniquely its own, thinking of new and unique solutions to issues in ways that Twitter failed to.
To suggest Mastodon has to solve its problems by being more like today’s Twitter is unimaginative and it fails to grasp what Mastodon is about.
Leave a Reply