Bluesky reached 40 million users on Friday and rolled out a "dislikes" feature to refine feed personalization, alongside new conversation controls designed to keep discussions civil without heavy-handed moderation.
Driving the news: The platform announced the milestone with updates including dislike-style feedback, improved toxic comment detection, and reply workflow changes—all aimed at letting users shape their experience rather than relying on centralized enforcement.
Why it matters: Bluesky is betting on user control over top-down policing. The platform's approach lets you train your algorithm to surface less of what you don't want—a philosophy rooted in its decentralized DNA that's sparked debate as it scales.
How it works: The "dislikes" beta feeds a personalization engine. Tap dislike, and the system learns to surface less of that content type in your Discover feed and reply rankings. Think of it as training wheels for your algorithm, without displaying public dislike counts.
The big picture: Bluesky's tackling a problem competitor Threads stumbled over: landing users mid-conversation with strangers. The platform's mapping "social neighborhoods"—clusters of people who regularly interact—to prioritize replies from folks closer to your circle.
- New toxic comment detection downranks spam, off-topic jabs, and bad-faith replies in threads and notifications.
- The Reply button now opens the full thread before you compose, nudging users to read before firing off redundant takes.
- Reply settings got more visible, so you can control who responds to your posts.
Between the lines: This follows a month of unrest as some users pushed for the platform to ban controversial figures outright. Bluesky's doubling down on its toolkit instead: moderation lists for quick blocking, content filters, muted words, and the ability to detach quote posts—a feature that undercuts X's toxic "dunking" culture.
Reality check: Decentralized moderation means users run their own show, but a vocal subset wants centralized enforcement. Bluesky's threading the needle by improving tools rather than wielding the ban hammer—a gamble that could define its identity as it scales.
What's next: The dislikes beta rolls out soon, with social neighborhood mapping and improved toxicity detection already in testing. Whether this approach keeps 40 million users engaged—and attracts more—depends on whether personalization beats out platform-level policing in the court of public opinion.





















