{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifsd5qjapswht52yq5ahyjvleftrlovusrzkysajd6ejyvuhhhtzy",
    "uri": "at://did:plc:pi6woz4d47bkuws673w2il2r/app.bsky.feed.post/3mleubmjrr3k2"
  },
  "path": "/t/anti-llm-sentiment-considered-harmful/14008?page=5#post_89",
  "publishedAt": "2026-05-08T21:19:35.000Z",
  "site": "https://discourse.haskell.org",
  "textContent": "I’m not sure why people are complaining about the tiresome effort of reviewing PRs when this is something that AI is unambiguously good at.\n\nYou should use AI to review all PRs, including those generated by AI. The value of doing this becomes immediately obvious once you start. Of course not all the issues and queries the AI raises will necessarily be valid or worth fixing, since it errs on the side of pointing things out. And human review is still helpful, but can at that point focus on overall direction and design rather than looking for fiddly mistakes. Trust me (or just try it), AI is MUCH better than you at finding fiddly mistakes, as well as a hundred other little nits that may not have even occurred to you.\n\nI think at some point all serious software projects are going to do this routinely.\n\nIn addition, you should have an LLM look over the PRs you intend to submit, just to save back-and-forth time.",
  "title": "Anti-LLM Sentiment Considered Harmful"
}