Current news should keep its subject

When you ask Mu for news about a specific topic, you get news about that topic. Not a summary of the day’s headlines. Not a paragraph that starts with what you asked for and drifts into a general briefing. The answer stays on the subject you chose, and it tells you how recent the stories are.

This matters because trust with news is fragile. A person asking about a local election wants to know what happened in that race, not a list of national political stories. Someone asking for market news on a particular sector wants the data for that sector, not the Dow Jones. When Mu’s news search cannot find recent articles on the exact topic, the answer leads with that caveat before sharing what it did find. The trade-off is plain: you see the limitation, not a smoothed-over version that pretends the information is fresh.

Mu does not broaden the query to fill space or make the answer look more complete. That kind of broadening is a common pattern in AI assistants, but it undermines the reason someone asked the question. The user’s intent — a specific subject, a specific recency — is preserved. The assistant is honest when it does not know, and it does not substitute generic headlines for the thing you actually asked for. That is the baseline for a tool that respects your attention.