One page, every story in AI and robotics that matters, each told once.
Official blogs of AI labs, technology press, Google News searches, Hacker News, Reddit,
Product Hunt, the outlets most cited in AI posts on X (Bloomberg, Reuters, FT, Nikkei…), robotics labs, GitHub releases and new repositories, and Hugging Face
(trending models, Spaces, datasets and daily papers). The full list is public in sources.toml.
Articles about the same event are grouped: first by matching titles and version numbers, then by an AI editor (Claude) that merges reworded duplicates and follow-ups, and a second pass that compares the last two days of stories with each other.
A story rises with the number of distinct sources covering it (official sources count 3x, major outlets 2x), its importance (1 to 5, judged by the AI editor) and its freshness (the score halves every 12 hours). The lead link goes to the official source when there is one, otherwise to whoever reported it first.
Open / Closed: open weights or open-source code versus proprietary. Official: the lead source is the company itself. New to try only lists things you can use, download or run today: no rumors, no plans.
Headlines and summaries are written by an AI from the sources' own text and can contain mistakes. Always check the linked source.