@schizanon

I'd like a feature (for Tusky) that shows me the fraction of posts (account's posts / all posts in my personal timeline) of each account I follow, for a specified timeframe, say last 1/3/6 months.

And ideally, there'd then be a feature to show a thinned out personal timeline that contains only toots of the more silent accounts (starting with the most silent) so that there's in total only x number of toots per week (say 70 or so).

#timeline #timelinehacking #algorithmictimelines

I doubt that most people defending the "simple reverse chronological timeline" in #Mastodon, opposing #algorithmicTimelines, have any idea of how it currently works.

Because, having debugged that code and the SQL, I can confidently say that hardly anyone can understand the current algorithms.

There's thousands (inc. dependencies) lines of code and logic, generating very complex SQL queries to fill your seemingly simple reverse chronological timeline. Complex algos already fill your timeline.

@gedeonm @Iconfactory

This reminds me of the anecdotal evidence that number of 'followers' and 'likes' with large follower accounts sometimes don't equal actual eyeballs as there are also some streaming services that have large followers numbers in the millions but views only in the thousands, oddly. Hmm.

#AlgorithmicTimelines

#funfact

#ARPANET1 routes around #ToxicSocialMedaOnePointOh #MLTimelnes #algorithmictimelines

Vs. #mastodon #ChronilogicalTimelines #nomlrequired

#funfact

#ARPANET1 routes around #ToxicSocialMedaOnePointOh #MLTimelnes #algorithmictimelines

Twitter is a microphone connected to a mixer owned and controlled by Elon Musk.

If you understand how microphones and mixers work, you also understand why you can’t “stay and fight.”

#twitter #fediverse #twitterMigration #algorithmicTimelines #BigTech #socialMedia #SiliconValley #capitalism #ElonMusk #billionaires #fascism

Looking at the #analytics of the old 🆓🐦 site makes me sad to see the day that in less than two months the sale of #Twitter made the site completely useless and the move to #web3 accelerate beyond the #ToxicSocialMedaOnePointOh of #AlgorithmicTimelines driven by #Meta #Instagram #Snapchat #WhatsApp & #Twitter dying a horrible race to the bottom #ARPU #DAU #DAP driven #decline of #web1 & #web2 technologies. 👎

☠️💀⚰️🥀R.I.P. 💀☠️🥀⚰️⚱️😐
👎☠️ #AlgorithmicTimelines ☠️👎

On the flip side: can we find a middle ground? Can we address the challenges some other way? If we do move toward #AlgorithmicTimelines can we find ways to circumvent the problems?

There’s a lot of rich potential in this design space and lots of room to experiment.

Side note: I hate the term “algorithmic timeline” because reverse chron sort is an algorithm, but here we are.

Objections to this concept of #AlgorithmicTimelines:

1. If I _am_ online a lot it can make it harder to catch up, not easier. I see the same things again and again and can’t scroll back.

2. It can create “engagement bubbles” (this happens anyways, but people trust them less and they look different).

3. If my use case for the platform is following friends with similar posting habits it can be really hard to catch up at all just by scrolling backwards.

Core reasons to not want time-ranked #timelines (or to want what people term #AlgorithmicTimelines):

1. If I’m trying to catch up on what’s happening since the last time I logged in then that’s not the best model unless I basically live online.

2. It lets the system auto-filter posts likely to be low quality in replies without removing them entirely.

3. At high following counts it means I don’t have to manually manage a ton of lists or curate my experience based on others posting behavior