# About highsignal.sh & Editorial Standards

highsignal.sh delivers calm, precise, and internet-native artificial intelligence news. Our mission is to help readers understand what matters in AI in five minutes, backed by trustworthy primary sources when you want to go deeper.

## Verification Pipeline & Evidence Grounding

Every story published by highsignal.sh undergoes strict evidence grounding:
1. **Multi-Stage Ingestion**: Monitored feeds from frontier AI research labs, academic publishers, and technical channels are bounded and sanitized against SSRF and XML attacks.
2. **Autonomous Evidence Validation**: Candidate stories must cite verified primary artifacts (technical reports, evaluation suites, open weights, official lab announcements) within a 72-hour freshness window.
3. **Verbatim Quote Checking**: Explanations must be grounded in retrieved evidence. Synthetic claims, hallucinated quotes, or uncorroborated benchmarks are strictly rejected by policy.
4. **Source Tiering**: Community discussions (e.g. Hacker News, Reddit, X) serve solely for story discovery and are never published as fact without corroborating primary proof.

## Editorial Independence & Transparency

- We do not accept paid placements, sponsored reviews masquerading as news, or compensated story rankings.
- AI models used in our curation pipeline have no network tools and are constrained strictly to retrieved evidence text.
- Material corrections are appended with transparent revision notes recorded directly on the published story. Report factual discrepancies to corrections@highsignal.sh.

## Reader Privacy

- Zero reader tracking: no advertising cookies, behavioral analytics, user profiling, or paywalls.
- Public reading endpoints are open and free to all human readers and automated AI agents.

## Core Links
- Front page: https://highsignal.sh/
- Contact & Newsroom: https://highsignal.sh/contact
- Privacy Policy: https://highsignal.sh/privacy
- LLMs Index: https://highsignal.sh/llms.txt