[jalev]

Take the first stories I found from this month's Clarkesworld[1] or Granta[2] or BCS[3] and read the prose. Notice the specificity of the language, how the doesn't try to insist upon itself? Notice how very few metaphors are actually in prose? Notice how, even when writing about fictional worlds and concepts, the language used grounds the _stories_ being told and not the concepts?

And then look at the submissions for unslop. This is the best we can get? Cliche-driven, over-metaphor'd, statistically-average purple-purpose _content_? It's sad, really, that we're many years into this entire thing and it still can't produce something that doesn't have my eyes drifting from the page.

[1] https://clarkesworldmagazine.com/khan_07_26/

[2] https://granta.com/here-comes-the-sun/

[3] https://www.beneath-ceaseless-skies.com/stories/the-ecstasy-...

[pseudo0]

Funny that you mention Granta, given their recent issue with allegedly AI generated work: https://www.theguardian.com/books/2026/jun/20/granta-magazin...

One thing to note here is that the Unslop contest rules required generating the entire work with a single prompt, no editing or modifications allowed. Those are pretty challenging conditions that leave a lot of improvement within easy reach, all it would take is an editing pass or two.

[LearnYouALisp]

Training data aka forums, marketing sludge, fan fiction, SEO scrapers (of SEO scrapers...), then the even broader lower-bar places that came after 00s

[sph]

> Cliche-driven, over-metaphor'd, statistically-average purple-purpose _content_

If this is expected from LLM generated prose, why don't we expect LLM generated code to exhibit the same qualities?

> It's sad, really, that we're many years into this entire thing and it still can't produce something that doesn't have my eyes drifting from the page

It's great. Human creativity is still king despite the attempts to reduce it to a few algorithms for talentless hacks to exploit with the click of a button.

Who but the sociopath would hope to supplant human creativity with a machine they control? I wish your position wasn't so widespread in these parts.

[solid_fuel]

> If this is expected from LLM generated prose, why don't we expect LLM generated code to exhibit the same qualities?

That's the fun part, it does! I think people who don't pay much attention to the code they ship don't see it, but LLM written code has a lot of the same problems that LLM written prose does. It's repetitive, muddled, and relies too much on crutches - constant boilerplate and pointless, inaccurate comments.

[InsideOutSanta]

What might be bad for prose (predictable, boring) might be desirable for code. Maybe that's why LLMs work well for writing things read by computers, but not so much for things read by people.

[lelanthran]

> Maybe that's why LLMs work well for writing things read by computers, but not so much for things read by people.

They don't really work well for that, though.

The reason you hardly ever hear about it is because the people delivering code via LLMs aren't critically evaluating the code it generates. This is why Claude Code, a text app that is little more than glue between various text-suppliers, is what, 500k SLoC in a high level language?

[InsideOutSanta]

> people delivering code via LLMs aren't critically evaluating the code it generates

I guess this is the flip side of "you're prompting it wrong": "you only think it works because you can't tell how much it sucks."

[pigpop]

Core vim is around 300k lines of C89 code and it's just a text editor.

It's over a million lines if you include all the scripting, language features and tests.

[sph]

Vim has been around for decades, supports a billion architectures and an entire modal editing system plus a scripting language. And it’s in C, not a high level language so verbosity is higher. It’s really not comparable to a glorified frontend to an API.

[pigpop]

It turns out things are more complex than they may initially appear.

[sph]

Unless you're writing enterprise Java, conciseness and simplicity of design is still the ideal to aim for; those are not the adjectives I would use to describe LLM generated code.

Laziness is a feature. When you have a tool that is the exact opposite and solves code problems with more code, all you have is a machine that generates tech debt at exponential pace.

[Xirdus]

If code is predictable then it should be extracted into reusable functions/classes/modules and reused in accordance to DRY principle. I'm not a fan of this AI future where coding standards drop to the floor because humans won't be reading that code anymore.

[InsideOutSanta]

Predictable and redundant are not the same thing. Also, DRY is not a hard rule. Applying DRY like it's a rule creates bad code.

[scotty79]

> Cliche-driven, over-metaphor'd, statistically-average purple-purpose

That's high literature for you. That's why so few people read it. Most prefer more down to Earth books, but AI doesn't default to that style.

The problem for AI might be that humans wrote very few good books. If you train a model for literary purposes you should weight training material by quality. Which is hard to evaluate.

> It's sad, really, that we're many years into this entire thing and it still can't produce something that doesn't have my eyes drifting from the page

Since internet happened, I have this problems with 98% of human written books. A book must have some very strong hooks to keep me reading till the end. "Blindsight" barely made the cut.

[GeoAtreides]

That's not what high literature is. That's like looking at some clever linux kernel code and dismissing it in favor of a small nodejs backend.

Good literature is difficult (not always, of course). Just like you can't go from a couch potato to running a marathon in one day, you can't jump from Brandon Sanderson to enjoying Gormenghast (or something like the The Worm Ouroboros). It's impossible. It takes effort, it takes time and it takes a lot of reading to appreciate what the real masters can do with mere words.

[scotty79]

If something requires effort to value it, is the value in the thing itself? Or is the thing garbage and all the value is in your effort?

[isomorphic_duck]

I didn’t think I would see this take on HN of all places. You can’t appreciate Higher Topos Theory before spending a decade’s worth of effort in pure math - does that make much of Modern Algebraic Geometry “garbage”?

[scotty79]

Math doesn't exist to be appreciated. Literature has literally no other application.

EUV litography machine is incredibly complex and I would have to learn for decades to meaningfully understand it, yet I can appreciate it knowing very little (of it and in general).

[GeoAtreides]

I climb mountains to see the views; the views are there independent of how much effort I put on the path (or how much satisfaction I draw at the end of trek), but I do have to put in the effort to see the views.

[msrp]

Would you take a free mountain lift ride to the top if you happened to come across one during a climbing trip?

[GeoAtreides]

would I read the summary of a book expecting the same thing as reading the book?

no

[jplusequalt]

>That's high literature for you.

What is "high literature"? Have you actually read any of the greats? I have, and while I'm not a fan of everything I've read, I never felt inundated with constant metaphors and overly eloquent prose.