Twitter Tests New Ways to Suggest Content, Including “Dreams”
(Twitter Tests Subconscious Recommendations, Dreams)
SAN FRANCISCO – Twitter is trying out new methods to recommend posts. These methods look at subtle user behaviors. The goal is to understand what people might like without them actively saying so. This is different from just following accounts or liking tweets.
The company calls this approach subconscious recommendations. It analyzes small signals. These signals include how long someone looks at a tweet. They also include tiny movements like hovering over a link. Twitter thinks these small actions show true interest. They might be better than obvious likes or retweets. People don’t always interact openly.
Alongside this, Twitter is testing a feature named “Dreams.” This feature suggests content during natural breaks in scrolling. If a user pauses reading for a few seconds, Dreams might show a relevant tweet. It aims to feel helpful, not pushy. The idea is to offer interesting posts when someone might be thinking about what to see next.
Twitter says user control is important. People can turn off these features. They can also adjust their settings. The company promises privacy is protected. The signals used are anonymous and aggregated. No individual’s detailed subconscious data is stored alone.
(Twitter Tests Subconscious Recommendations, Dreams)
Early tests are happening with a small group of users. Feedback will guide further development. Twitter wants its timeline to feel more personal and engaging. Finding posts users truly enjoy is key. The company hasn’t announced a full launch date yet. More testing is needed first. Engineers are working to refine the algorithms. They want the suggestions to feel natural and useful.