Thread

I can't tell who's teaching who anymore

34 points · 31 comments · eustoria

Loading the complete thread in the background. This saved snapshot is available now. Refresh

ben_w · 1790781950 · focus · HN ↗
Mm. Same on the em-dashes.
Hopefully one of the only ways I'm in a top-10-HackerNews-user list is #7 user of em-dashes before ChatGPT was relased: <a href="https://www.gally.net/miscellaneous/hn-em-dash-user-leaderboard.html" rel="nofollow">https://www.gally.net/miscellaneous/hn-em-dash-user-leaderbo...
I guess one irony, is that while there's many people here bemoaning LLMs as "nondeterministic", for prose in general it would be much better if they acted like such complainers meant, rather than the non-determinism being in such a narrow band as to be an instant cliché.
1. mcphage · 1790797854 · focus · HN ↗
  Nice! I only made it to #13 :-/
A_D_E_P_T · 1790782008 · focus · HN ↗
I know this feel. I was a fan of the em-dash long before the LLMs started spamming it. Now I am forced to use the semicolon more heavily.
genxy · 1790782012 · focus · HN ↗
This is performance art. Right in the middle of the "article" is this golden nugget.
> My Graveyard file sits right in the middle of this. Every entry in it is me editing my own writing because of something a model does. The same move those podcast hosts made with "delve", only pointed the other way.
1. jeremyjh · 1790782279 · focus · HN ↗
  Yeah, I was getting strong AI vibes before that but that was the clencher. I do believe it was heavily edited by a human, or maybe "humanized" by Astra.
Uhhrrr · 1790782054 · focus · HN ↗
teaching whom
1. c-b · 1790782264 · focus · HN ↗
  beat me to it
  1. genxy · 1790782314 · focus · HN ↗
    it is engagement bait
jdw64 · 1790782229 · focus · HN ↗
In my country, Korea, genre fiction communities are huge. So AI-related writing has become quite trendy lately.
In Korea, people call the AI writing style "쇠맛 the taste of metal.(or metallic taste)" They say it tastes like a machine.
When Koreans generally define the metal taste, they usually put it something like this: "The plot is terrible, but the sentences are really excellent and refined."
Generally, unless you're using a fine-tuned local model, sentences with that metal taste are considered refined. So there's this irony where people say writing not done by AI lacks refinement.
In fact, even very famous Korean authors have been caught using AI, but people didn't really notice—they just felt the prose had become very elegant.
Anyway, I'm rambling, but the thing is, AI sentences were basically "learned and erudite" by default. When I entered grad school in Korea, I was taught where to use hyphens, told these were "things you need to learn when writing papers," and learned language techniques while being taught how to write papers—and AI uses those techniques really well. So it's ironic.
It's just a shame that in order to avoid writing with that metal taste, you have to avoid the very vocabulary that used to be recommended.
1. oliculipolicula · 1790828729 · focus · HN ↗
  Ah.. starts to make a bit of sense.
  It could be like when I step into an unfamiliar cafe, I can sort of tell straightaway it's a Korean and not Japanese just by the "nonsymbolic" interior design.
  Would you say it's the opposite of wabi-sabi? That's supposed to "taste natural" I think. But raw.
  Seems like two opposite "cultural","statistical" decisions starting from the same commercial question
amelius · 1790782243 · focus · HN ↗
It should be simple to fit a style-filter onto an LLM. Curious why we still don't have that, and why everybody is using the default.
1. RGS1811 · 1790784809 · focus · HN ↗
  I think the idea that there could be such a thing as a “style” filter misunderstands what style is. It’s not about avoiding buzzwords, it’s about tone, empathy with a reader, the layered and often unconscious will to express certain ideas with a specific flavor, and the idiosyncrasies that come from being a particular person who has lived a particular life with its own odd mix of influences and experiences.
  Humans fail at good style more radically than LLMs do, and LLMs are on average better writers than people. But good style requires particularity and history in a way that a model could never achieve with just a filter.
  The closest I’ve gotten to getting an LLM to write well was when I gave it an intellectual biography along with its writing assignment. But even that fails over longer context, because attention makes it very hard to build up narrative (even just conceptual narrative) in a compelling way that extends past the length of a short blogpost. Past a certain point you always end up back in the uncanny valley with that weird lack of cohesion and the unjustified contrastives, tricolons, and empty cliches.
MitPitt · 1790782348 · focus · HN ↗
> I take out em dashes and replace them with commas or full stops
Yeah that's your problem. I like my em dashes. You don't need to be ashamed of your style just because it got mainstream. Doesn't matter whether it's AI or a fashion fad. It's like liking a music band before it gets popular on tiktok.
1. happytoexplain · 1790782991 · focus · HN ↗
  Well, no, not wanting your writing to be mistaken for AI is not the same emotion as not wanting your writing to be interpreted as having "mainstream" or "fad" elements. They're completely different motivations.
  1. MitPitt · 1790783758 · focus · HN ↗
    It's the same emotion as if you're enjoying ska music subculture, but can't dress in its style of clothes, because people will think you're a skinhead.
    You'll be judged for riding in your old Tesla car, because people may think you're a fan of modern Elon.
    It may sound like a stretch, but I feel it's virtually the same thing — style.
2. john_strinlai · 1790783008 · focus · HN ↗
  >You don't need to be ashamed of your style just because it got mainstream. Doesn't matter whether it's AI or a fashion fad.
  i think it is a pretty different scenario from liking a band before it gets popular or chasing a fashion fad.
  when your career (or passion) is sharing your writing, it feels shitty to immediately and constantly have people dismiss your efforts as ai. within a few minutes of this being posted, multiple commenters claimed this was ai and that the author was trolling.
  ive personally given up on sharing any of my writing publicly and stopped blogging because i just dont have the energy to deal with the misplaced hatred that internet commenters have turned up to 11.
3. gdulli · 1790783047 · focus · HN ↗
  If someone thinks my taste in music is lame because I like Dave Matthews Band, I'll wear that. But if someone thinks I don't care enough about the quality of my output to produce it myself, that's a more real and tangible hit to my reputation.
4. mcphage · 1790797747 · focus · HN ↗
  > You don't need to be ashamed of your style just because it got mainstream
  It's not about shame, it's about your work being dismissed because it's an indicator that you didn't put any effort into it.
  Although by the end, the article veers somewhat, and reads more like "I put a lot of effort into my writing, and just because I use AI also doesn't mean that I want to be lumped in with the people churning out AI slop". Which, okay, but that's not quite as noble a position as they seemed to start off with.
yipinwong · 1790782349 · focus · HN ↗
Inbred degrades DNA biologically.
Inbred stems from less variety.
If you hang out with the same person, you degrade in personal quality.
If you learn at the same spot, same places, same people all the time, your learning degrades.
Same for the AI, they inbreed and they degrade.
---
I purposefully hangout with people (not all the time still) outside of my circle for that exact reason.
jeremyjh · 1790782359 · focus · HN ↗
[dead]
1. john_strinlai · 1790782492 · focus · HN ↗
  it wouldnt be an HN thread without an immediate accusation of ai usage.
  1. archagon · 1790802285 · focus · HN ↗
    A community centered around curiosity and engaging conversation would not permit AI-generated articles in the first place.
  2. jeremyjh · 1790803451 · focus · HN ↗
    AI slop is also not curiosity-piquing nor will it facilitate good conversations.
    john_strinlai · 1790803907 · focus · HN ↗
    i've seen this exact comment before too.
    imagine how nice the front page of HN would be if we just flagged the ai-generated articles instead!
    jeremyjh · 1790806499 · focus · HN ↗
    Why not both?
2. layer8 · 1790782874 · focus · HN ↗
  I mean, the whole article is the author bemoaning that LLMs have usurped his writing style.
  1. jeremyjh · 1790784879 · focus · HN ↗
    Yes. I think its satire, or something wierder.
3. happytoexplain · 1790783212 · focus · HN ↗
  Good lord, the very topic of the article is how sad it is that people are now pointing to human writing tropes and saying "must be AI", because AI over-uses those tropes. Maybe AI was used, maybe it wasn't. But we could have a modicum of charity.
  I can't tell you how many times I've written "it's not X, it's Y", or "it's like X, but Y", or used em dashes over the years before AI. Probably thousands.
4. striking · 1790783722 · focus · HN ↗
  [delayed]
spockz · 1790782577 · focus · HN ↗
This hits a note. It has so many aspects.
People who made something their professional life goal to become good at something have their work marked as AI, and even worse, like effortless or slop. Just because a tool has been trained to match that good style or type of output. A good system (tailwind) is not used directly anymore because the LLMs take all the wind of their sails by embedding the knowledge in the model.
This happens all without any form of attribution or compensation.
It’s not nice. But I don’t know what to do about it.
drakonka · 1790782753 · focus · HN ↗
I used to be a big dash lover in my writing. Now I shamefully cull them as well. For my personal blog I don't care that much, but for university essays etc where they explicitly disallow AI formulation I just worry they'll get the wrong idea.
dhanush · 1790782894 · focus · HN ↗
I was considering adding a permanent disclaimer to every blog post on my site about it not being AI written. Then I relented. Because I don’t have to explain to strangers on the internet that I wrote it myself. I like writing, and even if it is imperfect at times I’m okay with that. But I do catch myself editing my own words multiples times a day. Because I'm less sure if that sentence structure is truly mine or a side effect of me conversing with LLMs so much.
I'm trying to counter this by trying to read more fiction, but for every page that I read, I'm probably reading 50 pages or so of LLM output. So I think this is a battle I'm going to lose out on in the long term. And I can probably only delay the inevitable.
But yeah, I don't like beginning to lose or doubt my own writing voice one bit.
1. ftnnn · 1790828518 · focus · HN ↗
  Slang and imperfection expose the ghost. Embrace it.
  Added ya blog to my RSS feed.
fwlr · 1790783134 · focus · HN ↗
I used to joke about a hypothetical programmer with so much mechanical empathy that he would oppose fuzzers because a program would find it very uncomfortable to be fuzz-tested.
It now occurs to me, non-humorously, that’s what this person and many others are going through with AI: it’s a multi-year fuzz test of their sense of self, their methods for ideation and expression, and their ethical principles on attribution and intellectual property.
FLeXMurphy · 1790783213 · focus · HN ↗
There are multiple subjects covered in this and it makes no sense. The first 3/4s deal with one subject - the reshaping of language, then the last 1/4 go to talk about Tailwind's layoffs, due to search engines/LLMs upfront providing answers.
The whole dashes debacle never made any sense, because many people are aware if you press and hold the hyphen "button" on your mobile touchscreen, it presents you options with different lengths you can choose from. That it was ever used as a signifier of LLM-written text meant that it can only show a false positive.
1. john_strinlai · 1790783361 · focus · HN ↗
  >then the last 1/4 go to talk about Tailwind's layoffs
  how are you measuring length?
  because tailwind is not mentioned until the very last paragraph, 1900 words into a 2000 word article.
  1. FLeXMurphy · 1790787396 · focus · HN ↗
    The article repeats and belabors over the first point, so I just quantize it. Simple as.
    john_strinlai · 1790792802 · focus · HN ↗
    it still seems exceptionally dishonest to claim that 3 sentences is 1/4 of the article.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.