{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiedr6k4h3alc74h3afs37tey4fmcg2zf5bqqgr6fhxo76gntqz6au",
    "uri": "at://did:plc:pi6woz4d47bkuws673w2il2r/app.bsky.feed.post/3mldfayen3ev2"
  },
  "path": "/t/anti-llm-sentiment-considered-harmful/14008?page=5#post_88",
  "publishedAt": "2026-05-08T07:59:15.000Z",
  "site": "https://discourse.haskell.org",
  "tags": [
    "SpringerLink",
    "ChatGPT is bullshit - Ethics and Information Technology"
  ],
  "textContent": "clinton:\n\n> This is slow. Not stupidly slow but Java slow (arguably worse as Java has had a lot more resources put into it’s optimising JIT compiler)\n\n“Slow” as in not appropriate for solving the class of low-level problems where running a garbage collector is an active hindrance, the area where C/C++/Rust excel. For the vast majority of real-world problems correctly executing the necessary sequence of steps with efficient data structures in mind is plenty fast as is, so having a language that can abstract all of the underlying complexity tidily is preferable.\n\nThat being said some parts of Haskell do absolutely suck performance-wise. I wish I could program on the type level same way I do on the term-level, the inability to do so makes both precompilation and mutable records unfeasible without introducing separate GHC extensions (something noone bothered with so far).\n\nclinton:\n\n> Also the term “hallucinate” is stupid. [..] The term “lying” is far more accurate\n\nThe correct term for it is “bullshitting”.\n\nSpringerLink\n\n### ChatGPT is bullshit - Ethics and Information Technology\n\nRecently, there has been considerable interest in large language models: machine learning systems which produce human-like text and dialogue. Applications of these systems have been plagued by persistent inaccuracies in their output; these are often...",
  "title": "Anti-LLM Sentiment Considered Harmful"
}