Our Sources

Every day we check 150+ hand-picked sources — the forums, research organisations, newsletters, and individual voices that shape how AI capability, safety, and governance actually move — and send you links to the few things that matter.

Beyond the list Wide net

The named sources below are the core, but they aren't the whole picture. Every edition also casts a wider net:

Broad web search. We run Google-style keyword searches across the open web each day, so a story breaking outside our regular sources still gets caught.
Trending topics. We track what's surging in the AI conversation and tag every item we publish — building a picture of which themes are trending over time.
Prediction markets. We monitor AI-related prediction markets daily; a sharp swing is an early-warning signal we investigate and surface (see the latest in the archive).

Government & policy

Forums & discussion

On X / Twitter

Official lab & company accounts plus the researchers, engineers, lab leaders, and policy voices we read.

Missing something?

This list keeps growing. If we're missing a source you trust, tell us — or just reply to any digest email.

The raw feed Firehose

Everything above, unfiltered. A separate mailing list that sends every new AI-related item our pipeline picks up — typically 50–150 links in one email every morning, with no curation, ranking commentary, or synthesis. It exists for people who want to pipe the raw links into their own tools and do their own filtering.

If you just want to stay informed, you almost certainly want the curated daily digest instead.