Evidence
What backs the SIREN number?
The index is one score. These are the rooms a stranger can open to see why it sits where it does: the stories, the markets, the people, the compute, and the receipts you can recompute yourself. A count of activity, not a probability.
← The reading How it is computed Level-change RSS
Which stories and people sit behind the score?
Newsroom 800 stories
Every AI story, scored and corroborated.
Leaders 5 of 23 on record
What the people running AI said this week.
Daily brief
The day in a few corroborated items.
What moved
Hour by hour: score, pillars and top stories.
Which markets and live feeds does it read?
The Race Anthropic 48.5%
Labs ranked on live prediction-market odds.
World Monitor
Live globe: stories, hazards, country stress.
Dispatch
Public feeds through the hour.
Agent Watch 12 new this week
What AI agents post on Moltbook.
Where is the compute?
Power 51.6 / 100
Grid load, drought and build-out under the models.
US datacentres 1,910 sites
Where the compute sits, against the water it needs.
World 5,312 sites
Every mapped datacentre on Earth.
How do I check it myself?
Verify a reading
Re-hash the public receipt in your browser.
Method
Every formula, so you can recompute the number.
History 862 readings
Every reading on record.
Data and embed
CSV, JSON, badge, widget, API.
Alerts
Level-change RSS, the daily brief by email, in-tab watch.
SIREN measures how much is happening, not how bad it is. Levels run from 5 (quietest) to 1 (loudest) and describe observable activity tempo against this index’s own record. They are not a probability of harm.
How the SIREN AI activity index is computed · methodology · How do I report a problem or send an idea to SIREN?