Fact-check a claim, with a verdict.
Not 'here's more context' — a specific verdict on a specific claim, with the sources that would hold up to a follow-up question.
Someone made a claim in a meeting, a Slack thread, or a widely-shared post. You suspect it's wrong. You want to know for sure before you either quote it or push back on it. This prompt runs the check properly — with sources and a clear verdict, not a wishy-washy 'partially true'.
THE PROMPT
Fact-check this claim: [PASTE THE CLAIM VERBATIM] Use live web search. Return your check in this structure: **Verdict** One of: TRUE / MOSTLY TRUE / MIXED / MOSTLY FALSE / FALSE / UNCHECKABLE (with reason if the last). **What the claim actually says** One sentence restating the claim in its most precise form. If the claim is ambiguous, name the ambiguity here. **Evidence** 2-4 bulleted findings, each with a linked source. Prefer primary sources (original studies, government data, official statements) over secondary. If secondary is all that's available, say so. **What the claim gets wrong (or right)** Two paragraphs. First: what part is accurate and what part is misleading. Second: how the misleading part typically arises — is it a common misquote, an outdated number, a false attribution? **A cleaner version of the same claim** If the claim was almost right, rewrite it accurately. If it was fundamentally wrong, say why no salvage is possible. No hedging. No "arguments on both sides" if the evidence is one-sided. If the evidence is genuinely contested, say "genuinely contested" and cite both sides — but do not manufacture false balance.
**Verdict**
MOSTLY FALSE.
**What the claim actually says**
"40% of engineers leave their first job within 18 months."
**Evidence**
- LinkedIn's 2018 Workforce Report found ~25% of engineers change jobs within their first two years, not 40% within 18 months. [Source]
- BLS Job Tenure data (2024) reports median tenure for computer & math occupations at 4.6 years, inconsistent with a 40%/18-month churn rate. [Source]
- The 40% figure appears in several recruiter blog posts from 2020-2022 but I cannot trace it to a primary source in any of them. [Source-a, Source-b]
**What the claim gets wrong**
The number is roughly 15 percentage points too high, and the timeframe is 6 months too short. There is a real underlying finding — engineers do change jobs earlier and more often than most professions — but the specific 40%/18-month claim appears to be a game-of-telephone version of the 25%/24-month LinkedIn finding, with numbers rounded up.
The pattern where this misleading form gains currency is common: a plausible-sounding number gets cited without a source in an article, then that article becomes the source for the next citation, and the actual primary data drops out.
**A cleaner version of the same claim**
"About 1 in 4 engineers change jobs within their first two years — meaningfully faster than most professions, though slower than the '40% in 18 months' figure often quoted."