Bluesky adds Saved Posts feature
Bluesky added a new “Saved Posts” feature that is practically similar to what other Social Media platforms have been having for a few decades. Their announcement thread also mentioned that there is a new way for existing users to move their 📌 emoji marked previous “bookmarks” to actual bookmarks with a new pin2saved web application. Essentially that both improves people’s own privacy by reducing unneeded previous data sharing and also improves the overall status of Bluesky as a platform, by cleaning up the frequently annoying 📌 posts from the timelines.
Feature announcement was visible via in-app notification overlay, with the following text (also visible on the screenshot):
“Saved Posts
Finally! Keep track of posts that matter to you.
Save them to revisit anytime.”
On the web UI of Bluesky, Saved Posts feature is now available via “Saved” side menu item. It is empty initially, but starts to fill up after people start bookmarking posts with the newly added feature. List-based UI of the new feature is nicely minimalistic at the moment, but it might get more complicated eventually. One potential future improvement might be a way to add Saved Posts filtering to enable faster searching from the list of past bookmarks. That would improve usability after people have started to add increasingly large numbers of bookmarks via the Saved Posts feature.
While Saved Posts don’t make any major difference for most users, the key change is also on how Bluesky is moving backwards to a more centralized data storage. Even while they promoted themselves as a “distributed” platform with the self-developed AT Protocol data syndication, in practice they are largely centralized. Saved Posts feature improves usability of the service, but also makes it more clear that distributed platforms have limitations. While each of the social media posts on Bluesky gets broadcasted via their APIs, there are increasing amount of demands for improving the safety of people using the service.
Sadly it feels like that people might wake up to the sad reality too late, realizing that their posts had already ended up to 3rd party databases for unlimited number of use cases. There aren’t much stopping external developers from using Bluesky posts as a training data for machine learning and LLMs, and people might end up to bad situations when that happens more in the upcoming future.
Technically each of the Bluesky user accounts is a small SQLite-based database repository that holds both the public posts and past activity feeds of an user. Tools exist for fetching that repository for analysis, and such functionality enables external parties to do analytics and detailed analysis on how people have used the platform. Latest change is moving part of the data away from the public view, thereby improving the privacy, but at the cost of losing public visibility to a part of the previous activity data contents. While noise to regular users, 📌 posts added more visibility to what items other people viewed important enough to (potentially) read again later.
