{
  "id": 68538,
  "title": "Test 2.0 - yet another farce - hand labelling danger 2.0",
  "url": "/competitions/airbus-ship-detection/discussion/68538",
  "author_name": "Alexander Veysov",
  "post_date": "2018-10-14T06:21:38.305000",
  "votes": 7,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Without further ado, it looks like the hosts actually labeled only &lt;3k new images with ships for the new test dataset.</p>\n\n<p>This looks like a farce to me. Such labeling can realistically be done in one-two days for a moderate fee. Seriously, doesn't it sound ridiculous?</p>",
  "messages": [
    {
      "id": 403627,
      "postDate": "2018-10-14T06:21:38.307Z",
      "content": "<p>Without further ado, it looks like the hosts actually labeled only &lt;3k new images with ships for the new test dataset.</p>\n\n<p>This looks like a farce to me. Such labeling can realistically be done in one-two days for a moderate fee. Seriously, doesn't it sound ridiculous?</p>",
      "rawMarkdown": "Without further ado, it looks like the hosts actually labeled only &lt;3k new images with ships for the new test dataset.\n\nThis looks like a farce to me. Such labeling can realistically be done in one-two days for a moderate fee. Seriously, doesn't it sound ridiculous?",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "403627": "Without further ado, it looks like the hosts actually labeled only &lt;3k new images with ships for the new test dataset.\n\nThis looks like a farce to me. Such labeling can realistically be done in one-two days for a moderate fee. Seriously, doesn't it sound ridiculous?"
  }
}