This site is a reading list, not a newsroom. Headlines are pulled from public RSS feeds published by technology outlets, research labs and AI companies. Every item links to the original publisher; the words and images remain theirs.
What is here
The news page is a rolling view of model releases, research results, infrastructure and policy. The explainers are longer pieces about how the technology actually works — attention, inference economics, small models and reading research without getting lost.
The feed is refreshed every 30 minutes. If a source is unreachable, the previous edition stays online rather than showing an empty page.
Sources
TechCrunch, The Verge, WIRED, MIT Technology Review, OpenAI, Google AI, Hugging Face, arXiv cs.AI, MarkTechPost and The Decoder. The mix is deliberately small: a handful of outlets that cover both products and the research underneath them.
Editorial rules
- Link to the source. Always.
- Keep summaries short and descriptive, not sensational.
- Do not republish full articles or images.
- Prefer primary sources — papers and lab posts — when they exist.
Privacy
No analytics, no advertising scripts, no third-party fonts. The site is a directory of static files behind nginx. The only outside requests happen when you click a headline and leave for the publisher.
AI news moves fast; understanding moves slowly. This site is built for the second speed.
Aggregated headlines are provided for discovery and commentary. All rights belong to the original publishers. If you operate one of the feeds and want it removed, the feed will be dropped on request.