About Signal Grid

Signal Grid is an AI news intelligence hub that aggregates stories from multiple sources, clusters them into events, and scores each event to help you understand what matters in AI.

How It Works

1. Ingest — Pollers continuously fetch stories from RSS feeds, Hacker News, Reddit, and ArXiv.
2. Cluster — Stories covering the same underlying event are grouped using a multi-signal composite of title similarity, keyword+IDF weighting, entity overlap, temporal proximity, and domain diversity.
3. Score — Each event cluster is scored on coverage breadth, recency, engagement, and velocity to produce an overall “bigness” score.
4. Present — Events are displayed with full provenance receipts so you can trace every claim back to its source.

Scoring Methodology

• Coverage Breadth (25%) — How many independent sources cover this event.
• Recency (25%) — Exponential decay based on time since the most recent story.
• Engagement (35%) — Trust-weighted community engagement across sources.
• Velocity (15%) — Breaking-news detection based on story accumulation rate.

Sources

Currently tracking MIT Technology Review, The Verge, Ars Technica, VentureBeat, TechCrunch, Hacker News, Reddit, and ArXiv.