Supra 𝔖

X made two interesting algorithm updates in October that i think content creators should pay attention to and neither of them is about simply getting more likes according to X's public algorithm repository: October 8: the Phoenix ranking model improved how it predicts time spent on posts October 13: Phoenix retrieval was updated to help surface posts people read all the way through this doesn't mean there's some guaranteed hack where longer posts automatically perform better but it does reinforce something i've been saying for a while: your content needs to be worth consuming here's what i'd focus on: >make the first two lines interesting enough to earn attention >give people a clear reason to continue reading >replace generic observations with practical examples >use numbers and evidence where they genuinely add value >break complicated ideas into digestible sections >remove anything that doesn't improve the explanation >make educational posts useful enough to revisit >test different formats instead of assuming text always wins >share actual experiments rather than repeating popular opinions >measure whether people engage beyond simply seeing the post >build content around problems your audience actually has >create stronger endings that deliver on the opening promise and something i'd be careful about: don't confuse making content longer with making it better a 150-word post that explains one idea perfectly can be more useful than a 1,000-word post full of unnecessary information the objective isn't to force people to spend more time reading it's to make that time worthwhile source: github.com/xai-org/x-algorit…