AI Panic Academy
A transparent guide to how AI Panic collects reporting, reduces noise, measures attention, creates evidence-based Signals and connects news with market context.
AI Panic organizes information and highlights observable patterns. It does not guarantee accuracy, predict outcomes or provide investment advice.
01
AI Panic brings configured public news sources and feeds into one readable intelligence stream. Items are normalized so titles, publishers, timestamps, links and categories are handled consistently. Related coverage is grouped around the same event, preventing syndicated stories from looking like separate breakthroughs. Ranking uses freshness, importance and coverage activity to surface meaningful developments. It prioritizes reading; it does not certify truth. Every item keeps its publisher, original link and timestamp.
02
AI Panic is mainly an aggregator and intelligence layer, not the original publisher. Source names and original links remain visible. Near-identical articles are clustered; copies of one syndicated story should not resemble independent confirmation. Several sources matter when they add distinct reporting or evidence. Inclusion is not endorsement: publishers, changing stories and automated classification can be wrong. Important claims should be checked against originals, dates and other reliable reporting. Missing data and uncertainty are shown rather than invented.
RelatedSources·FAQ·Hallucination glossary
03
The Attention Score measures unusual verified company coverage against its normal level. Recent unique coverage is compared with a historical 28-day baseline excluding the latest 24 hours, so a spike cannot inflate its own comparison. Near-identical stories become one event. Recency, source diversity, event activity and history contribute to the 0–100 score, with components exposed. Insufficient history is labeled. Attention measures coverage intensity—not sentiment, company quality, price direction or trading advice.
RelatedAttention Score glossary·Signals·FAQ
04
Signals neutrally indicate that a defined condition occurred. Categories are Coverage Spike, Breaking Cluster, Sustained Attention, Unusual Market Movement, Price/News Divergence and Multi-Source Confirmation. Each shows evidence, a plain-language explanation and story links. Stable event identities suppress repetition. News Signals can continue without market data, while market-dependent Signals pause when required data is stale. Signals describe observations; they do not prove causation or predict returns.
05
AI Market Pulse connects a curated AI-company universe with news intelligence. The current plan stores daily snapshots, not continuous real-time quotes. One bundled Marketstack attempt is permitted on eligible US trading days after the New York market-close slot; no weekend requests or repeated daily retries. Cards show currency, basis, snapshot date, update time and stale/unavailable states. Charts use local history. Conservative matching adds Market Context only for clear company references. News near a price move does not prove causation. This is not investment advice.
RelatedAI Markets·FAQ·alerts
06
Accounts add a personal workspace. Supported AI-50 companies can be watched from lists, charts and matched articles. Alerts monitor price targets or daily percentage movement, fire once when a condition becomes true, re-arm after it leaves that state and pause with stale data. Saved articles and preferences do not change editorial ranking. Free and Plus allowances come from one catalog. Downgrades retain data but may pause excess alerts; upgrades do not silently reactivate them. Available and Coming soon Plus benefits remain separate.
More questions? The FAQ covers the product in Q&A form.