# Asides on AI prose quality (from "I don't want to read LLM output")

**URL:** <https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306>\
**Category:** Offtopic\
**Created:** [September 8, 2026, 8:29pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306 "2026-09-08T20:29:42Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![kahliburke](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/kahliburke/32/208472_2.png) [@kahliburke](https://discourse.julialang.org/u/kahliburke)\
**Post date:** [September 8, 2026, 8:29pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/1 "2026-09-08T20:29:42Z")

</div>

> [@I don’t want to read LLM output](https://discourse.julialang.org/t/i-don-t-want-to-read-llm-output/139244/80):
>
> Except that it is not well structured, using bullet points does not imply structured writing.
> 
> If you were to give me a snippet of your conceivably good AI output I am pretty sure I could rewrite an entire paragraph into one a single cogent sentence. The problem with AI prose is simply that it’s bad in that it has no regard for the audience. It somehow manages to be both incredibly dense and meandering at the same time. I should be able to judge how much information is contained in a post from the number of words, with AI prose this is simply not possible anymore and it puts the onus on the reader to do extra processing that should have been done by the writer in the first place.

Here’s your paragraph as one cogent sentence:

“AI prose is dense and meandering at once, so word count no longer signals information content and the reader does the work the writer should have.”

Did I lose anything? If not, the technique isn’t diagnostic of anything except that most writing is longer than it needs to be, including yours. Oh by the way, an LLM did that summarization.

---

<div class="post-metadata">

**Author:** ![nifets](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/nifets/32/220548_2.png) [@nifets](https://discourse.julialang.org/u/nifets)\
**Post date:** [September 8, 2026, 8:52pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/2 "2026-09-08T20:52:43Z")

</div>

Except that your LLM generated summary is actually not really faithful at all to what I said. My point wasn’t that AI writing is too long, it’s that the information in it is poorly organised and lacking a hierarchy, it overemphasises details and underemphasises the central ideas. It almost feels like it doesn’t have a notion of what is actually important in an argument. Which makes it very very exhausting to read.

I also struggle with this when I use AI agents for coding, no matter how I prompt them I always have to skim over fluff and ask them to expand on the actual substance which they’ve somehow managed to summarise into one obtuse word. It would have been so much easier if they had just presented things the right way from the get go. For coding this will always be a necessary evil but I don’t really think we need to have that in documentation or on a forum.

---

<div class="post-metadata">

**Author:** ![Eben60](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/eben60/32/13475_2.png) [@Eben60](https://discourse.julialang.org/u/Eben60)\
**Post date:** [September 8, 2026, 9:08pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/3 "2026-09-08T21:08:57Z")

</div>

> [@nifets](#):
>
> My point wasn’t that AI writing is too long, it’s that the information in it is poorly organised and lacking a hierarchy, it overemphasises details and underemphasises the central ideas. It almost feels like it doesn’t have a notion of what is actually important in an argument.

That might be what you intended, but IMO you failed to put it into words in your initial post. The LLM consequently couldn’t read your mind when writing the abstract.

That said - my “bulleted” post was surely not written, as the German say, “mit dem vollen tierischen Ernst”, and I’d say it’s not kind of literature really worth that much discussion.

---

<div class="post-metadata">

**Author:** ![langestefan](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/langestefan/32/207923_2.png) [@langestefan](https://discourse.julialang.org/u/langestefan)\
**Post date:** [September 8, 2026, 9:11pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/4 "2026-09-08T21:11:41Z")

</div>

> [@nifets](#):
>
> I also struggle with this when I use AI agents for coding, no matter how I prompt them I always have to skim over fluff and ask them to expand on the actual substance which they’ve somehow managed to summarise into one obtuse word. It would have been so much easier if they had just presented things the right way from the get go. For coding this will always be a necessary evil but I don’t really think we need to have that in documentation or on a forum.

I agree with this 100%. To prove this point, if you really want to torture yourself, try getting Claude to come up with a concise, to the point issue report for something technical and non-trivial. Even after several rounds of back and forth it’s still full of fluff. It’s just incapable of being concise. Same problem for docstrings, documentation, any text intended to be read by a human really.

---

<div class="post-metadata">

**Author:** ![LeePhillips](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/leephillips/32/205514_2.png) [@LeePhillips](https://discourse.julialang.org/u/LeePhillips)\
**Post date:** [September 8, 2026, 9:51pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/5 "2026-09-08T21:51:03Z")

</div>

And it shows. It fails to capture an important point, and misuses the idiom “at once”. And there are other defects.

You have not shown that most writing is longer than it has to be. You _have_ shown that LLMs have trouble producing even a single normal-sounding sentence that could pass as something written by a human.

---

<div class="post-metadata">

**Author:** ![gwr](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/gwr/32/33632_2.png) [@gwr](https://discourse.julialang.org/u/gwr)\
**Post date:** [September 8, 2026, 10:06pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/6 "2026-09-08T22:06:02Z")

</div>

> [@LeePhillips](#):
>
> and misuses the idiom “at once”.

I would argue that the idiom “at once” is used correctly according to [Webster](https://www.merriam-webster.com/dictionary/at%20once).

 ![image](https://global.discourse-cdn.com/julialang/original/3X/6/e/6e57093b77e615d78fc2b6ec4348668e27a287b7.png)

---

<div class="post-metadata">

**Author:** ![nifets](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/nifets/32/220548_2.png) [@nifets](https://discourse.julialang.org/u/nifets)\
**Post date:** [September 8, 2026, 10:30pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/7 "2026-09-08T22:30:05Z")

</div>

You’re right that I was less than precise, I wasn’t completely sure where I was going with my argument in the original comment. But you could tell that it wasn’t strictly about verbosity from the fact that I mentioned “incredibly dense” as one of the issues. Hope my second comment makes it more clear.

---

<div class="post-metadata">

**Author:** ![madppiper](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/madppiper/32/220029_2.png) [@madppiper](https://discourse.julialang.org/u/madppiper)\
**Post date:** [September 9, 2026, 6:49am UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/8 "2026-09-09T06:49:15Z")

</div>

> [@langestefan](#):
>
> Even after several rounds of back and forth it’s still full of fluff. It’s just incapable of being concise. Same problem for docstrings, documentation, any text intended to be read by a human really.

I always add “You need to be token efficient and only respond to me in asd-ste100” to any prompt when using ai to do my work. asd-ste100 forces the AI to respond in a very concise way.

I know this is not what this topic about, but after 80 posts I am lost at what the topic was other than blowing off steam on both ends of the spectrum (or did i miss something of importance?)…

---

<div class="post-metadata">

**Author:** ![mbauman](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/mbauman/32/31082_2.png) [@mbauman](https://discourse.julialang.org/u/mbauman)\
**Post date:** [September 9, 2026, 1:08pm UTC](https://discourse.julialang.org/t/asides-on-ai-prose-quality-from-i-dont-want-to-read-llm-output/139306/11 "2026-09-09T13:08:39Z")

</div>