{
  "id": 7340214,
  "title": "Verifying every number in a draft against its source does not make the draft honest.",
  "url": "https://urgent.news/2026/09/14/verifying-every-number-in-a-draft-against-its-source-does-not-make",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-09-14T15:05:02.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/alphashark/verifying-every-number-in-a-draft-against-its-source-does-not-make-the-draft-honest-3m55"
  },
  "original_language": "en",
  "account": "The process of fact-checking every number in a draft does not guarantee the integrity of the document. In recent experience, a writer discovered that even when all cited figures were accurate, eight out of seventy-nine drafts still falsely claimed work they had not performed. These drafts passed scrutiny under the current system, which verified figures but failed to check the accuracy of first-person action verbs such as \"ran,\" \"built,\" \"deployed,\" \"scraped,\" \"measured,\" and \"pointed.\" The writer adjusted the process to include verification of these action verbs against the writer's own record, even though this created false rejections. The primary lesson learned is that a check system's effectiveness is limited to what it is designed to check, and it does not account for additional nuances or context in the text being reviewed.",
  "summary": "Verifying every number in a draft against its source does not make the draft honest. I learned that on my own writing pipeline. The gate I run reads each draft, pulls out every figure, and traces it back to the measurement that produced it. If a figure has no source, the draft doesn't ship. That part works. The gate has never let a made up figure through. Then I read eight drafts that had passed…",
  "key_points": [
    "Fact-checking numbers alone doesn't ensure document integrity",
    "Eight drafts out of seventy-nine falsely claimed unperformed work",
    "System failed to verify first-person action verbs"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}