Tracking AI existential risk. Source-backed context. Original reporting always linked.
MONITORING CORE AI-RISK FEEDS · UPDATED HOURLY
About PDOOM

Understand AI risk without drowning in the firehose.

PDOOM turns public reporting, research, policy, and expert forecasts into a source-backed view of what changed and why it matters.

Choose your depth

Where to start

The term

What does p(doom) mean?

p(doom) is shorthand for someone's estimated probability that advanced artificial intelligence causes human extinction or another irreversible global catastrophe. Estimates vary dramatically because the evidence, timelines, and assumptions remain disputed.

The purpose

What PDOOM does

PDOOM monitors a focused set of AI-risk, research, policy, security, and frontier-lab sources. It groups related coverage, highlights major movement, and keeps the original reporting one click away.

Read with context

What the numbers mean

Doom Meter

A composite discourse signal based on selected prediction markets and recent headline language. It is not a scientific estimate of extinction risk.

Attention score

A triage tool combining recency, source breadth, report volume, and topic relevance. It answers “what should I inspect first?”

Source count

A measure of coverage breadth. Multiple sources do not automatically mean that the sources agree or independently verified a claim.

Essential

A high relevance designation. It means the material is worth attention, not that PDOOM endorses its conclusions.

Coverage boundary

Focused, attributable, and inspectable.

PDOOM is not a general technology-news site and does not reproduce full articles. It surfaces material related to catastrophic risk, alignment and control, frontier capabilities, governance, misuse, corporate safety, and forecasting.

Every monitored story links to its original source. Scores help organize the evidence; they do not replace reading it.

The project

Built and maintained by Zac Boring.

PDOOM is an independent project for making AI-risk information easier to follow. Questions, corrections, or source suggestions are welcome.

Contact Zac →