{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigkpihwy2bhsdb42pil7ntw6p5avfruublagowgf2n5jj34rytn6m",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mjmcl23tou62"
  },
  "path": "/t/humans-see-it-instantly-llm-doesn-t/174765#post_4",
  "publishedAt": "2026-04-16T11:21:15.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "Great question! It’s hard to be 100% sure, but here are a few quick ways to spot AI-written replies:\n\n  * They often sound too polished, generic, or repetitive.\n  * They might lack personal experience, specific details, or follow-up questions.\n  * If the reply solves a complex problem in perfect textbook language without edge cases, it’s more likely AI-generated.\n\n\n\nAt the end of the day, a bit of healthy skepticism never hurts — even human replies can be misleading sometimes!",
  "title": "Humans see it instantly — LLM doesn’t"
}