Turning the news into your own feed: US app pushes ultra‑personalized news timelines

A US‑based news recommendation app, similar in spirit to services like SmartNews, has rolled out a major update in the last couple of days, turning its interface into something that looks and feels much more like a social media timeline.

Instead of a static list of “top stories”, the app now builds a personalized stream of news cards based on what you read, where you are, and even what time of day you tend to open the app. Behind the scenes it blends traditional machine‑learning with large language models, so it can adapt not just to topics you like, but also to how you like to consume them—for example, long explainers at night versus quick headline summaries in between meetings.

To reduce the classic “echo chamber” effect, the new engine also injects a small but steady dose of stories that present different angles or opposing views on topics you follow. The team says their ambition is to create a feed that nudges people toward a slightly wider perspective, instead of locking them into a narrow bubble. For Japanese readers who want a low‑friction way to keep up with English‑language news, this kind of app could become a handy second home screen.