Bluesky uses an algorithmic feed called “Discover” and your Home feed to show posts it thinks you will like. But the algorithm needs signals from you to learn your preferences. Without these signals, the feed shows random or popular content that may not match your interests. This article explains how likes and reposts act as training data for the Bluesky algorithm. You will learn exactly which actions influence the algorithm and how to use them to shape your feed.
Key Takeaways: Train the Bluesky Algorithm with Likes and Reposts
- Like button (heart icon): Tells the algorithm you want to see more content from that author and on that topic.
- Repost button (two arrows): Signals strong endorsement and increases the reach of that post to your followers.
- Settings > Feeds > Discover Feed: The main algorithmic feed that learns from your likes and reposts over time.
How the Bluesky Algorithm Uses Likes and Reposts
Bluesky does not publish the exact ranking formula for its algorithmic feeds. However, the platform has confirmed that likes and reposts are the primary signals used to personalize the Discover feed and the Home feed. Each like you give adds weight to the author and the topic of the post. Each repost carries even more weight because it shows you want the post to appear in your followers’ feeds too.
The algorithm also considers recency. A post that gets many likes and reposts quickly will rank higher than an older post with the same number of interactions. This means your actions help the algorithm decide not only what content to show but also when to show it. Over time, the algorithm builds a profile of your interests based on the cumulative likes and reposts you have made.
The Blobcat algorithm, which powers the Discover feed, uses a collaborative filtering approach. It groups users with similar like and repost patterns. If you like posts about photography, the algorithm finds other users who also like photography posts and surfaces their content to you. This is why consistent liking and reposting matter: they train the algorithm to place you in the right interest group.
Steps to Train the Bluesky Algorithm with Likes and Reposts
Training the algorithm requires deliberate action over several days. The algorithm does not respond instantly to a single like. You must build a pattern of behavior. Follow these steps to guide the algorithm toward your preferred topics.
- Like posts from authors you want to see more of
When you find a post that matches your interests, click the heart icon below the post. Do this for at least five different authors in the same topic area. The algorithm interprets this as a signal that the topic and the author are relevant to you. - Repost high-quality content that fits your niche
Click the repost icon (two arrows forming a square) next to the like button. Choose “Repost” from the menu. Repost only posts that strongly represent the kind of content you want in your feed. Reposting random posts dilutes the signal. - Mute or unfollow accounts that post off-topic content
If an account you followed starts posting about things you do not like, unfollow it or mute it. Use the three-dot menu on the post or profile and select “Mute account” or “Unfollow.” This removes negative training data from the algorithm. - Use the Discover feed daily to reinforce your preferences
Open Bluesky and switch to the Discover feed. Like or repost at least three to five posts in that feed every day. The algorithm uses your Discover feed interactions as the main training set for your interest profile. - Create custom feeds for focused interest groups
Use the search bar to find feeds like “Photography” or “Tech News.” Pin these feeds to your Home tab. Interacting with posts in custom feeds trains the algorithm to recognize those topics as high priority.
Common Mistakes When Training the Bluesky Algorithm
Liking too many posts indiscriminately
If you like every post you see, the algorithm cannot distinguish between content you truly enjoy and content you merely glanced at. Be selective. Like only posts that you would want to see again. A ratio of one like for every ten posts you read is a good starting point.
Reposting without reading the full content
Reposting a post based solely on its headline or image can train the algorithm to show you misleading or low-quality content. Always read the full post before reposting. The algorithm treats a repost as a strong signal, so make sure the post genuinely aligns with your interests.
Ignoring the mute and block features
The algorithm learns from what you ignore as well as what you interact with. If you see posts from a topic you dislike but do nothing, the algorithm may interpret your silence as neutral or even positive. Use the mute feature on accounts or keywords that consistently show content you want to avoid. Go to Settings > Moderation > Muted Words to add specific terms.
Expecting immediate results
The Bluesky algorithm updates your interest profile gradually. It may take three to five days of consistent liking and reposting before you notice a change in the Discover feed. Do not expect the feed to transform after a single session. Patience and consistency are key.
| Item | Likes | Reposts |
|---|---|---|
| Signal strength | Moderate | Strong |
| Effect on author ranking | Increases | Increases more |
| Effect on topic ranking | Increases | Increases |
| Visible to followers | No | Yes |
| Best use case | Daily feed tuning | Highlighting key posts |
If the Algorithm Still Shows Unwanted Content
If after a week of consistent liking and reposting you still see irrelevant content, try resetting your interactions. Go to Settings > Moderation > Muted Words and add the topics you want to avoid. Also, unfollow any accounts that post mixed content. Finally, check your custom feeds. If you have pinned a feed that does not match your interests, unpin it by clicking the gear icon on the Home tab and removing the feed. The algorithm will then focus on the remaining feeds and your interaction history.
You can now train the Bluesky algorithm to show content that matches your interests by using likes and reposts deliberately. Start by liking three to five posts in the Discover feed each day and reposting only the most relevant content. For advanced control, create custom feeds for your favorite topics and mute keywords that you want to avoid. The algorithm will refine its recommendations within one week of consistent use.