# Anti-LLM Sentiment Considered Harmful

**URL:** <https://discourse.haskell.org/t/anti-llm-sentiment-considered-harmful/14008>\
**Category:** Uncategorized\
**Created:** [April 30, 2026, 1:58pm UTC](https://discourse.haskell.org/t/anti-llm-sentiment-considered-harmful/14008 "2026-04-30T13:58:37Z")\
**Posts on this page:** 1\
**Showing post:** 70

<div class="post-metadata">

**Author:** ![srid](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.haskell.org/srid/32/5514_2.png) [@srid](https://discourse.haskell.org/u/srid)\
**Post date:** [May 6, 2026, 1:49pm UTC](https://discourse.haskell.org/t/anti-llm-sentiment-considered-harmful/14008/70 "2026-05-06T13:49:39Z")

</div>

I posted the below comment (reporting on how I use LLMs) on another thread, without realizing that this one existed:

> [@How to filter out vibe-coded dependencies](https://discourse.haskell.org/t/how-to-filter-out-vibe-coded-dependencies/13918/79):
>
> I use Opus to generate almost all of my code these days … bounded, however, by [structural review](https://kolu.dev/blog/hickey-lowy/) enabled by [srid/agency](https://agency.srid.ca/).
> 
> The last few PRs in [Emanote](https://emanote.srid.ca/) are all generated by Claude Code.

I’m clearly on the pro-LLM side, but I can relate to one point from the anti-LLM site — which is the review overhead. When a contributor opens a PR that was LLM-generated, why should I accept it when I could have done the same myself (especially near-autonomously using `srid/agency`?). I don’t know the answer to this; an idea I came up with was to encourage contributors to submit elaborate proposals instead (see [kolu](https://github.com/juspay/kolu/blob/master/CONTRIBUTING.md) and the [example](https://github.com/juspay/kolu/pull/760) PR), which—once finalized, based on human discussions—can be fed directly to LLMs. The contributors contribution will be the proposal & spec that still gets record in the repository.

And LLMs do generate ‘average’ quality code. Not to mention, [architecturally](https://kolu.dev/blog/hickey-lowy/) weaker one. And they are not creative enough to identify opportunities to extract frameworks or libraries out. All of this still requires the human in the loop, even though I believe it is possible to automate some of this in the future.

---

_[View the full topic](https://discourse.haskell.org/t/anti-llm-sentiment-considered-harmful/14008)._
