{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreianxi2de6nr3o7elcrnfwlqzwgsroztiudfu6jjzut7t5sna7c3wm",
    "uri": "at://did:plc:pi6woz4d47bkuws673w2il2r/app.bsky.feed.post/3mkufted33xk2"
  },
  "path": "/t/anti-llm-sentiment-considered-harmful/14008?page=3#post_45",
  "publishedAt": "2026-05-02T08:50:08.000Z",
  "site": "https://discourse.haskell.org",
  "textContent": "Circa 2015 a bunch of cryptocurrency nerds jumped on the Haskell bandwagon with the expectation that a terse language with the maximally strong type system will automatically produce good software. That turned out to be a thorough disappointment, as the software they wrote didn’t look functional and the confusing (to them) strictness rules tanked performance. They moved on to Rust, expecting it fix all of Haskell’s garbage collector “issues”, and probably found themselves disappointed with having to shove `Arc`s everywhere and with how verbose the language is. I imagine by now they’re bidding on LLMs to fix everything, all I can do is wish them luck.\n\nNone of what I’ve seen from Snoyman looks inspiring or rigorous, it’s all “good enough to do the task at hand” kinda code. Most of it would probably look almost the same in Rust, I don’t see much of an issue with him using tools he finds more appropriate for the job.",
  "title": "Anti-LLM Sentiment Considered Harmful"
}